Ashton-Tate dBASE III Reference Notes These dBASE III Reference Notes were obtained from the Ashton-Tate Support Library on CompuServe (GO ASHTON). Steve Mullen [71676,124] Menu 1 >>> @...GET...RANGE RANGE is used in conjunction with @...GET to specify an acceptable continuous set of input values to date or numeric fields. The RANGE is initialized by specifying lower and upper bounds (inclusive) which may be literal values, memory variables, or expressions. For example: 1) Literal values: @ 10,10 GETvar1 RANGE 1,9 @ 11,10 GET mdate RANGE CTOD('12/12/84'),CTOD('12/12/85') 2) Memory variables: STORE 1 TO low STORE 9 TO high @ 10,10 GET var1 RANGE low,high STORE CTOD('12/12/84') TO low_date STORE CTOD('12/12/85') TO high_date @ 10,10 GET mdate RANGE low_date,high_date 3) Expressions: @ 10,10 GET var1 RANGE low+365,high+(365*10) @ 11,10 GET mdate RANGE DATE(),high_date+120 Entries outside of the defined range will generate an error message and input will be prompted until a valid entry is made. >>> @...GET and READ (1) If an attempt is made to @...GET memvar PICTURE "999.99" and the memory variable is not initialized with the number of decimal places specified in the PICTURE clause, the GET display will not show the decimal digits for input. Entering a decimal point will exit the GET. For example: * ---Set up numeric input. num = 0 @ 10,10 SAY "Enter number " GET num PICTURE "999.99" READ * * ---After entering a decimal point during the READ, * ---the GET will display the following: Enter number : 0. : To properly input numeric memory variables with decimal digits using @...GET...PICTURE, initialize the memory variable with the number of decimal digits used in the PICTURE clause. For example: * ---The following command line assigns num with * ---the same number of decimal digits as found * ---in the PICTURE clause. num = 0.00 @ 10,10 SAY "Enter number " GET num PICTURE "999.99" READ (2) When using @...GET and READ cetain keys will terminate the READ. The following table shows which keys do so and where in the pending GETs the key will terminate the READ. Key: The READ will terminate at: ----------- --------------------------- Backspace First character of first GET Left arrow First character of first GET Right arrow Last character of last GET Up arrow First GET Down arrow Last GET Esc Anywhere F1 Anywhere Ctrl-[ Anywhere Ctrl-Q Anywhere Ctrl-W Anywhere Ctrl-Home Anywhere Ctrl-End Anywhere PgUp Anywhere PgDn Anywhere Ctrl-PgUp Anywhere Ctrl-PgDn Anywhere Alt 0-9 Anywhere >>> @...GET...PICTURE Assuming a memvar is initialized with zero (that is, memvar = 0), when the user types a period at an @...GET memvar PICTURE "9999.99", dBASE will proceed to the next command line. This will not occur when the memory variable is initialized to two decimal places (that is, memvar = 0.00). Only the integer portion of the number just entered will be stored to memvar if the period is typed. >>> @...GET...PICTURE The function "@B" is used with the @...SAY...GET statement to left-justify variables of numeric type. It is most useful when SAYing a numeric field or memory variable that is more easily understood as a character string, such as a part number. Use of this FUNCTION with GET, however, causes a slight change in the way numeric variables are read from the screen that may cause some difficulties. A numeric memory variable will default to a length of ten digits when initialized; however, if you are using the PICTURE function "@B" in an @...GET statement, a numeric memory variable will GET the width of the memory variable exactly as initialized, even if a template symbol is used. Initializing a memory variable to zero will cause the GET statement to display one zero on the screen. A READ command will allow one digit only to be entered to the memory variable. This occurs whether the memory variable is initialized to 0 or 0000. For example: SET DELIMITERS ON w = 1234 x = 0 y = 0000 @ 9,0 GET w PICTURE "@B9999" @ 10,0 GET x PICTURE "@B9999" @ 11,0 GET y PICTURE "@B9999" will produce: :1234: :0: :0: A READ command will allow four characters to be entered in the first variable, but only one character in the next two variables. If the "@B" function is used, initialize the memory variable to the proper length or the input may be truncated. >>> @...SAY...PICTURE To display a dollar-sign character ("$") in front of a numeric value and not have the possibility of a lot of "$"s filling the blank areas, do the following: * ---To display a single "$". STORE 123.56 TO num @ 10,10 SAY "$" @ 10,11 SAY num PICTURE "99,999.99" This will generate: $ 123.56 * ---The other option available is: STORE 123.56 TO num @ 10,10 SAY num PICTURE "$$,$$$.$$" This will generate: $$$123.56 >>> @...SAY using relative addressing We use the '$' function in dBASE II to control relative screen addressing with the @...SAY function. For example, you can have a command file print the contents of a datafile to the screen as follows: * ---This is dBASE II syntax. USE DO WHILE .NOT. EOF * @ 5, 5 SAY @ $+1,$ SAY @ $+1,$ SAY @ $+1,$ SAY SKIP * ENDDO [while .not. eof] The dBASE III utility, dCONVERT, is used to convert dBASE II programs, datafiles, etc. to dBASE III formats. dCONVERT does not change the '$' function to ROW() and COL(); it leaves it alone. However, this will not cause any problems when executing the code in dBASE III. dBASE III will treat the '$' function as a relative addressing function. >>> APPEND FROM [SDF/DELIMITED [WITH ]] The DELIMITED form of the APPEND FROM command should be documented as having a WITH clause. WITH is not mentioned in the reference section. Below are a few examples: Example 1. To read a comma delimited file in which the character strings are enclosed in double quotes: APPEND FROM DELIMITED or APPEND FROM DELIMITED WITH " Example 2. To read a comma delimited file in which the character strings are enclosed in single quotes: APPEND FROM DELIMITED WITH ' Example 3. To read a comma delimited file in which the character strings are not enclosed at all: dBASE III CANNOT READ A FILE OF THIS FORMT! Also, the syntax of the APPEND command does not include a WHILE option as the manual indicates. The correct syntax is: APPEND FROM [FOR ] [SDF/DELIMITED [WITH ]] APPEND FROM SDF expects records in the text file to be terminated with a carriage return/line feed (0DH 0AH) pair, in this order. If the order is reversed (0AH 0DH), dBASE III ignores the character following the carriage return. This causes the first character of every record (after the first record) to be missed in the resultant database file. Menu 2 >>> Closing Database Files Do not swap data disks without first closing all the open files on the original disk. Not doing so will cause the loss of data in buffers, and will also write a new entry in the directory of the new disk. USE will close the currently SELECTed database file and CLOSE DATABASES will close all database files in all work areas. CLEAR ALL also closes all open database files and will also release all memory variables. QUIT closes all files, releases all variables, and exits dBASE III. Choose one of these commands to close your files before swapping data disks. >>> COPY TO [SDF/DELIMITED [WITH ]] (1) COPY TO DELIMITED does not enclose logical fields with the specified delimiters. Numeric and date fields are also treated this way. Date fields go out in the format YYYYMMDD. (2) COPY TO DELIMITED WITH , encloses the character fields in commas and separates the fields with another comma. This command behaves differently from dBASE II as shown below: In dBASE II: . USE file1 . COPY TO file2 DELIMITED WITH , will result in: SANTA,CLAUS,NORTH POLE,ALASKA In dBASE III: . USE file1 . COPY TO file2 DELIMITED WITH , will result in: ,SANTA,,,CLAUS,,,NORTH POLE,,,ALASKA, >>> COPY FILE TO The COPY FILE command copies files in 512-byte blocks whereas the COPY TO command will copy a .DBF file until the end-of-file. Therefore, the COPY FILE command will usually create a slightly larger file than the COPY TO command. However, the COPY FILE is faster. >>> COPY STRUCTURE EXTENDED CREATE FROM COPY STRUCTURE EXTENDED and CREATE FROM are fully implemented in dBASE III although not documented. A brief description is given below. 1) COPY STRUCTURE EXTENDED creates a file in whch the field names become the contents of one record. The syntax for this COPY option is: COPY STRUCTURE EXTENDED TO 2) CREATE FROM forms a new database (.DBF) file in which the structure is determined by the contents of a file created with COPY STRUCTURE EXTENDED. The syntax is: CREATE FROM >>> CREATE/MODIFY REPORT When you get the error message "Internal error - bucket overfilled" while in CREATE REPORT or MODIFY REPORT, you have too many characters in the report. The maximum number of characters, or bucket size, is 1,440 bytes. This includes the number of characters in the following list: Report heading - plus one byte for each line Subtotal heading(s) - plus one byte for each line Subtotal expression(s) - plus one byte for each expression Field heading(s) - plus one byte for each line Field expression(s) - plus one byte for each expression The extra byte is a null terminator for each expression and heading. When there are multiple lines in a heading, dBASE III separates them with a semicolon in the .FRM file. >>> Date conversion from dBASE II The dBASE BRIDGE manual (pages 23-24) lays out an elaborate scheme for converting dBASE II "dates" to dBASE III date fields. A much easier way is to simply convert the dBASE II database file to a dBASE III file and modify the structure from character to date field. All dates stored in a dBASE II character field as "MM/DD/YY" will directly convert to a dBASE III date field. >>> Dates that are blank CTOD() and DTOC() are intended to be inverse functions. That is, if DTOC(date) = char, then CTOD(char) = date. This is true in all circumstances except when the date is blank and the character string is " / / ". To detect a blank date, you must use the DTOC() function rather than CTOD(). For example: reg_date = CTOD("11/09/84") ? reg_date = CTOD("11/09/84") .T. ? DTOC(reg_date) = "11/09/84" .T. * ---With a blank date the following occurs: blank_date = CTOD(" / / ") ? blank_date = CTOD(" / / ") .F. ? DTOC(blank_date) = " / / " .T. As is evident from the example, the blank date is handled differently than the non-blank date. >>> Debugging tip The RETURN and CANCEL commands will release all PRIVATE memory variables when returning to the dot prompt. This can make debugging difficult with versions 1.0 and 1.1 as you may want to check the status of variables used by the program. The following debugging utility can be set up as a command file. This utility, called Dot.PRG, will allow you to interactively enter commands, such as LIST MEMORY and LIST STATUS. You can insert the command "DO Dot" at various problem points in your program. When this command is encountered, you will be able to enter interactive commands without destroying the current environment of the program at runtime. Once the program is fully debugged, you can remove the "DO Dot" command lines. * Dot.PRG DO WHILE .T. ACCEPT 'DOT ' TO command IF LEN( TRIM(command) ) = 0 EXIT ENDIF &command ENDDO RETURN >>> Demonstration Disk (RunTime+) In the Developer's Release of dBASE III, the dBRUN programs from the Demonstration Disk are not compatible with the full system. Code crunched with DBC.COM from the Demonstration Disk can only be run with dBRUN from the Demonstration Disk. Code crunched with DBC.COM from the Developer's Disk can only be run with the full dBASE III system or the dBRUN disk, purchased separately. The error message: No database is in USE. Enter filename: is a common indicator that the incorrect dBRUN or dBC is being used. >>> DISPLAY and LIST The DISPLAY and LIST commands are documented inthe manual as not having a FIELDS clause as part of the syntax, while the ASSIST and HELP menu options assume the FIELDS clause is required. dBASE III will accept either syntax for these two commands. >>> Duplicate Keywords in Config.DB If two or more COMMAND = lines are contained in the Config.DB file, only the last COMMAND will execute. This is true of any duplicate = entry in the Config.DB file, with the exception of DELIMITERS, which can legitimately have two entries. This is true for versions 1.0 and 1.1 of dBASE III. >>> FILE() function The FILE() function only searches the current directory. SET PATH TO does not affect this function. If other directories are to be searched, they must be supplied to the function. For example, * ---This will not find Data.dbf, if Data.dbf is in the * ---subdirectory ACCTS. SET PATH TO \DBASE\ACCTS IF FILE( "DATA.DBF" ) DO Process ENDIF Workaround: * ---This method will work. mpath = "\DBASE\ACCTS\" SET PATH TO &mpath IF FILE( mpath + "DATA.DBF" ) DO Process ENDIF >>> CTOD() Function The dBASE III Reference Manual does not clearly state that the CTOD() function will convert invalid dates to valid dates. The day is adjusted first. ? CTOD("02/29/85") 03/01/85 <------------ Extra day is added to month. ? CTOD("02/29/84") 02/29/84 <------------ Leap years are correct. If the month is invalid, it is divided by twelve and the year is adjusted. For example: ? CTOD("13/01/84") 01/01/85 <------------ Extra month is added to year. ? CTOD("12/32/84") 01/01/85 <------------ Extra day is added to month, extra month is added to year. ? CTOD("99/99/62") Menu 3 >>> FIND / SEEK FIND and SEEK are both used to move the record pointer of an indexed database to the first instance of the index key that matches the search argument. FIND searches on a literal character string while SEEK searches on an expression the value of which may be character, date, or numeric. The proper choice of command is related to the context and data type of the index key. Generally, FIND will only be used to search for a literal character string and SEEK for all other searches. The following are some typical cases: (1) You have an index key that is character and are working from the dot prompt: . FIND Lee or . SEEK "Lee" (2) You have an index key that is numeric or date and are working from the dot prompt: . SEEK 1250 . SEEK CTOD('12/12/85') (3) You are working within a command file and are initializing a memory variable as a search key: STORE SPACE(10) TO skey @ 10,10 SAY 'Enter value to search for' GET skey READ SEEK skey (4) You have a database field that is character, Code, and the contents are numeric digits and right-justified: ACCEPT 'Enter code to search for' to skey SEEK SPACE(LEN(Code) - LEN(skey)) + skey (5) You are working with several databases and want to search for a key value in the current work area using a field variable from a non-active area with an alias name: SELECT 1 USE File1 INDEX File1 SELECT 2 USE File2 INDEX File2 SELECT File1 SEEK File2->Field1 >>> Function Keys F1 toggles the cursor control menu on and off in the following full screen edit modes. APPEND EDIT MODIFY LABEL MODIFY STRUCTURE BROWSE CHANGE MODITY REPORT >>> Get Diskspace In the Developer's Release use the DISKSPACE() function to get the amount of space left on the currently logged drive. The DISKSPACE() will return the number of free bytes on the default drive as a numeric value. dBASE III versions 1.0 and 1.1 do not have a function to return the amount of space left on the default drive. So, to get the amount of diskspace in these versions, use the PC/MS-DOS utility CHKDSK and import the results into dBASE III. The basic algorithm is as follows: 1. Create or have available a general purpose database file called Util.DBF. Util.DBF has one field called Util_line which is character type and has a length of 80. This database file will be useful for any of these kinds of survey operations into PC/MS-DOS. 2. RUN CHKDSK including the designator of the drive for which you want the space statistic for, and pipe the result into a text file entitled Util.TXT. Piping is a PC/MS-DOS capability that allows the results of a program to be sent into a text file. It is very useful for passing parameters between programs when there is no formalized interface. The syntax is: > ^_____ DOS piping symbol For more information on this capability, consult your PC/MS-DOS reference manual. 3. APPEND the text file, Util.TXT, into the database file, Util.DBF. 4. Assign to a memory variable the number of free bytes on the specified drive from Util.DBF. This operation requires that you GOTO the record that contains the free disk space information and then extract the number of bytes from the field using the SUBSTR() function. The following is a LIST of Util.DBF with the results of a CHKDSK report. When APPENDed into a database file, the first and last records are always blank. Records 2 through 6 contain statistics about the currently logged disk drive. Note that this is the currently logged DOS drive and not the DEFAULT drive SET in dBASE III. Records 8 and 9 contain statistics about the memory configuration of your computer. The number of bytes for each attribute of the drive and memory occupy positions 1 through 9 in the database field, Util_line. Record# 1 2 9965568 bytes total disk space 3 155648 bytes in 4 hidden files 4 90112 bytes in 22 directories 5 6000640 bytes in 397 user files 6 3719168 bytes available on disk 7 8 524288 bytes total memory 9 122480 bytes free 10 The code that will get the the number of free bytes on the specified disk drive is as follows: SET SAFETY OFF RUN CHKDSK > Util.TXT USE Util ZAP APPEND FROM Util SDF GO 6 diskspace = STR( SUBSTR( Util_line, 1, 9 ), 9 ) USE SET SAFETY ON RETURN >>> Get Current Directory dBASE III has no facility to get the name of the current directory. To get it you must RUN the PC/MS-DOS command CD and import the results into dBASE III. The basic algorithm is as follows: 1. Create or have available a general-purpose database file called Util.DBF. Util.DBF has one field called Util_line which is character type and has a length of 80. 2. RUN the PC/MS-DOS command CD, piping the result into a text file entitled Util.TXT. 3. APPEND the text file Util.TXT into the database file, Util.DBF. 4. Assign to a memory variable the name of the current DOS directory from Util.DBF. The code that will execute this algorithm is as follows: SET SAFETY OFF RUN CD > Util.TXT USE Util ZAP APPEND FROM Util SDF currdir = TRIM( Util_line ) SET SAFETY ON RETURN >>> Get Last Update Date and Time To get the date of last update for the currently SELECTed database file in the Developer's Release of dBASE III, use the LUPDATE() function. LUPDATE() returns the date of last update as a value of date type. dBASE III versions 1.0 and 1.1 currently do not have a function that returns the date of last update for the SELECTed database file. To get the date of last update in these versions of dBASE III, use the PC/MS-DOS command DIR, and import the results into dBASE III. The basic algorithm is as follows: 1. Create or have available a general purpose database file called Util.DBF. Util.DBF has one field called Util_line which is character type and has a length of 80. 2. RUN the PC/MS-DOS command DIR with the name of your database file, piping the result into a text file entitled Util.TXT. 3. APPEND the text file Util.TXT into the database file Util.DBF. 4. Assign to a memory variable the date of last update from Util.DBF for your database file. The code that will get the last update of the currently SELECTed database file is as follows: SET SAFETY OFF RUN DIR .DBF > Util.TXT USE Util ZAP APPEND FROM Util SDF lupdate = SUBSTR( Util_line, 25, 8 ) luptime = SUBSTR( Util_line, 34, 6 ) USE SET SAFETY ON RETURN >>> INPUT The INPUT command does not initialize a memory variable of any type if is pressed at its prompt. The dBASE III manual says this will produce a syntax error (page 4-58). What really happens is that a syntax error will result if the non-existent memvar is later referenced. >>> MACRO (&) Substitution in a Format (.FMT) File Macro (&) substitution will not work in a format (.FMT) file. For example: Structure for database: Macro.DBF Number of records: Date of last update: 05/01/85 Field Field name Type Width Dec 1 FLD01 Character 10 2 FLD02 Character 10 3 FLD03 Character 10 ** Total ** 30 * Program..: Macro.PRG * Note.....: This program will not display any field from * the database file. CLEAR PUBLIC month month = " " USE Macro ACCEPT "Enter two digits 01, 02, or 03: " TO month SET FORMAT TO Macro READ CLOSE FORMAT CLEAR RETURN Macro.FMT contains one line: @ 05,30 SAY Fld&month If the SET FORMAT TO Macro, READ, and CLOSE FORMAT command lines are removed and DO Macro.FMT is inserted in their place, the program will execute as expected and the field will be displayed. >>> MEMO fields (1) MEMO fields are used to contain up to 5,000 characters of text information that is to be associated with a database record. Information may be read into a MEMO field using Ctrl-K-R and written to text files using Ctrl-K-W. Information from MEMO fields can be displayed or printed by using LIST, DISPLAY, ?. The field must be specified with these commands. However, these commands cause the MEMO field to wrap at 50 columns. The REPORT FORM may be used to output MEMO fields with line widths of more or less than 50 characters. (2) PACKing a database file with memo fields will not decrease the amount of disk space used by the .DBT file. The command file below demonstrates how to remove the deleted records and free the unused disk space. SET DELETED ON USE Filea COPY TO Temp CLOSE DATABASE ERASE Filea.dbf ERASE Filea.dbt RENAME Temp.dbf TO Filea.dbf RENAME Temp.dbt TO Filea.dbt SET DELETED OFF (3) When APPENDing FROM a database file with a memo field, the target .DBT file will be at least as large as the source .DBT regardless of how many records are APPENDed. This is because the target .DBT file contains all the information that was contained in the source .DBT file. Be sure to note this any time you are moving records from one database file that contains memo fields to another. There is a danger of unexpectedly filling a disk when doing this operation. >>> MODIFY STRUCTURE In MODIFY STRUCTURE, Ctrl-Home will bring up a menu on line 0 with the four choices listed below. Bottom Top Field # Menu They are selected with cursor control keys and . Bottom moves the cursor to the last field, Top to the first. Field # allows selection of a field number, then moves the cursor to it. Menu toggles the cursor control menu off and on. This feature is not documented under MODIFY STRUCTURE. >>> MODIFY STRUCTURE The "Warning" in the documentation on page 4-73 of dBASE III version 1.00 and WARNING of page 4-80 of dBASE III version 1.10 should read as follows: WARNING: Although you may change field names and field lengths, if you change both at once, the data of the fields that have been modified will not be appended into the new structure. Note that in the ASSIST mode, the following screen message is not entirely true, "Information in the database file is preserved where field names remain the same." As noted above, if only field names are changed or only the length, the data is appended into the new structure. The correct procedure when both field name(s) and field length(s) need to be changed, is to modify the field name(s) first, and then re-enter MODIFY STRUCTURE and modify the field length(s). Also, note that deleting a field (with Ctrl-U) has the same effect as modifying the field length. Therefore, make deletions at the same time you make field length changes. >>> LABEL FORM command If TRIM() functions are the only expressions on a line in a LABEL FORM, dBASE III will return "*** Execution error on SUBSTR(): start point out of range" when the LABEL FORM encounters a record in which all TRIMmed fields are blank. >>>LOCATE and CONTINUE The LOCATE and CONTINUE commands cannot be used between two or more work areas at the same time. If a LOCATE is issued in one work area then issued again in another work area, a CONTINUE in the original work area will result in a SKIP to the next record, not a CONTINUE. The following program segment illustrates how to LOCATE and CONTINUE between two data files. SELECT 2 USE Test SELECT 1 USE Test1 LOCATE FOR One > 1 DO WHILE .NOT. EOF() ? "FIELD Test1->One: " , One SELECT 2 LOCATE FOR one > 1 DO WHILE .NOT. EOF() ? "FIELD Test->One: ", One SKIP LOCATE NEXT 1000 FOR One > 1 ENDDO ^-------------- should be larger SELECT 1 than last record SKIP number in file. LOCATE NEXT 1000 FOR One > 1 ENDDO The dBASE III Reference Manual entry for MODIFY LABEL, states that "when you place several field names on one line of the LABEL contents screen, dBASE III eliminates the extra blank spaces and prints the label with only one space between fields." This should read, "when you place several field names, separated by commas, on one line, etc." Concatenations of fields appear as they always do. Menu 4 >>> Numeric fields with decimal places Although not documented, dBASE III expects the user to allow for a leading digit and a decimal point on numeric fields containing decimal places. For example, a numeric field of two decimal places should not be defined any smaller than four digits in length -- one position for the leading digit, one position for the decimal point and two positions for the two decimal places. If the structure for a numeric field does not allow for a leading digit (such as a width of three and two decimal places), numeric input to the numeric field will always be stored as zero. Also, if other numeric fields follow this field, they might automatically be zeroed out when numeric data is entered to the first field. >>> Numeric input of large numbers If a variable is initialized to zero, dBASE III does not allow input larger than 10 digits when using the @...GET command, even if the PICTURE clause is used. For example: x = 0 @ 5,5 SAY "Enter digits" GET x PICTURE "99999999999" * (There are eleven 9's) ----------^ READ The display is: Enter digits 0 If an eleven digit value is entered, the display is: Enter digits ********** (10 asterisks) This can be avoided by initializing 'x' to a value greater than ten digits (such as, 1000000000). This problem does not occur if a field is used rather than a memory variable. >>> PARAMETERS, passing Fields In the documentation concerning PARAMETERS when used in conjunction with the DO [WITH ] command, page 4-76 of the version 1.0 manual states, "A passed parameter may be any legitimate expression." Also, in the Glossary (page 7-3) the definition for Expression is, "Expression may consist of a field, a memory variable, a function, a constant, or any combination thereof." However, when a DO is invoked with a field in the parameter list, dBASE III will give the message, "Variable not found." In order to use a field name in the parameter list, you must use the Alias -> Fieldname form. For example: USE Filea DO WITH Filea -> Field1 will work, but the following will not. USE Filea DO WITH Field1 Menu 5 >>> PRIVATE In dBASE III, all variables are PRIVATE to the routine in which they are initialized unless otherwise declared. Variables created at the dot prompt will automatically be PUBLIC no matter how they are declared. Declaring a variable PRIVATE in a command file hides any outer-level definition of a variable with the same name from the current routine. It also hides any deeper-level routines from viewing any outer-level definition of a variable with the same name. In the example below, programs B.PRG and C.PRG do not have access to variable X in program A.PRG. Program C.PRG, therefore, will display X with a value of 15 and not 10. Program A.PRG, however, will diplay X with the old value of 10 even after executing program B.PRG. * A.PRG x = 10 DO B ----------> * B.PRG ? x PRIVATE x RETURN x = 15 DO C ----------> * C.PRG RETURN ? x RETURN In programming, you will want to declare a variable PRIVATE in a subroutine if you do not want this variable to interfere with an outer-level variable having the same name. To illustrate the use of PRIVATE, the command files MAIN.PRG and SUB.PRG are listed below, with the displayed output. Notice that all the variables are released when MAIN.PRG returns control to the dot prompt. Also notice that the variables initialized in MAIN.PRG are PRIVATE in the memory display even though they are never explicitly declared. Lastly, notice that the value assigned to height in SUB.PRG is not returned to MAIN.PRG, but the value assigned to area in SUB.PRG is returned. This is because height is declared PRIVATE in SUB.PRG and area is not. LISTINGS: * MAIN.PRG * -------- area = 0 height = 304 ? "Before call to SUB:" ? "-------------------" DISPLAY MEMORY DO Sub ------------------> * SUB.PRG ? "After call to SUB:" * ------- ? "------------------" PRIVATE height DISPLAY MEMORY height = 30 RETURN area = 10 * 20 * height * EOF: MAIN.PRG ? "Inside SUB:" ? "-----------" DISPLAY MEMORY RETURN * EOF: SUB.PRG OUTPUT: Before call to SUB: ------------------- AREA priv N 0 ( 0.00000000) HEIGHT priv N 304 ( 304.00000000) 2 variables defined, 18 bytes used 254 variables available, 5982 bytes available Inside SUB: ----------- AREA priv N 6000 ( 6000.00000000) HEIGHT priv (hidden) N 304 ( 304.00000000) HEIGHT priv N 30 ( 30.00000000) 3 variables defined, 27 bytes used 253 variables available, 5973 bytes available After call to SUB: ------------------ AREA priv N 6000 ( 6000.00000000) HEIGHT priv N 304 ( 304.00000000) 2 variables defined, 18 bytes used 254 variables available, 5982 bytes available lines will display: 0 variables defined, 0 bytes used 256 variables available, 6000 bytes available >>> PROCEDURE -- Calling Command Files from Procedures within dBASE III To call a command file from a procedure, you must follow a few rules. Rule 1: The command file cannot have the same name as any of the procedres in the file even if the extension is included as part of the filename. An attempt to do this will cause the inappropriate error message, "Unrecognized phrase/keyword in command." For example: * Proc_ne.PRG PROCEDURE One * ---The next command will not work because * ---Two is a procedure in this file. DO Two.PRG RETURN * PROCEDURE Two * ---The next command will work. DO Three.PRG RETURN * EOF: Proc_One.PRG This can be avoided by renaming either the command file or procedure. To avoid this problem you might want to begin procedure names with a prefix that command files will not have. For instance, in the previous example the procedures could have been called P_One and P_Two. Rule 2: Once the command file is invoked from a procedure file, it must not DO another procedure in the procedure file. Instead, it should RETURN to the calling procedure. Otherwise, the called procedure will usually execute, but an error message will be displayed for a command line that does not exist. Rule 3: Internal procedure calls (that is, a procedure that calls either itself or another procedure in the same file) must be kept to eighteen nested calls or less. The nineteenth call attempt will return execution to the calling command file with no error message. >>> PROW() When issuing SET PRINT ON, PROW() will be set to 2, regardless of its previous setting. An EJECT will reset PROW() to 2, not 0. Issuing an EJECT with SET PRINT OFF will reset PROW() to 0. When SET PRINT is ON, PROW() may never equal 0 or 1. Any attempt to test for PROW() = 0 or PROW() = 1 will work only when SET PRINT is OFF. Menu 6 >>> PUBLIC PUBLIC is used to declare memory variables as global and to prevent their release when control is returned to the dot prompt. PUBLIC variables must be declared prior to being initialized, and once declared, these variables will be assigned a logical false value until initialized. PUBLIC variables can be re-declared as PUBLIC without losing the values already stored in them. In programming, declaring all variables as PUBLIC in the main routine would make dBASE III behave similar to dBASE II. However, there is one difference in dBASE III, PUBLIC variables can only be released by the CLEAR MEMORY, CLEAR ALL, and RELEASE commands, but not the RELEASE ALL command. >>> Ramdisks There are several options for users who wish to use a ramdisk in combination with dBASE III. 1. For faster operation of applications that utilize routines stored in the DBASE.OVL file, you may wish to put the .OVL in a ramdisk. (a) The minimum drive size will have to be in excess of 181,000 bytes. The DBASE.OVL file for version 1.1 is 180,736 bytes. The total amount of RAM in your machine must be more than 440,000 bytes in order to do this. Additionally, if you are using a CONFIG.DB, it must be present on the drive where .OVL resides. (b) Boot dBASE III from the ramdisk by calling for the DBASE.COM from the drive on which it resides. For example: drive D: is the ramdisk and the DBASE.COM is on the C: drive. C> D: D> C:DBASE 2. It may be a very useful area for procedure or command files to be run from, increasing the speed of processing. Prior to entering dBASE III, copy the appropriate files to the ramdisk. Once in dBASE III, SET the DEFAULT TO the ramdisk drive and proceed. 3. It is also useful as a small work area to manipulate utility and temporary files. The useage tips on getting the current directory or diskspace are good examples of where a small ramdisk would be extremely useful and time efficient. >>> RELEASE, incorrect syntax The syntax "RELEASE ALL LIKE ," is incorrect but will not produce an error message. The correct syntax is either RELEASE or RELEASE ALL LIKE . You may use wildcard characters in the RELEASE ALL LIKE form of the command, but dBASE III will ignore any after the first comma. >>> RELEASE ALL, CLEAR MEMORY RELEASE ALL and CLEAR MEMORY are not equivalent commands as the dBASE III manual states. CLEAR MEMORY clears all memory variables, regardless where they were initialized. RELEASE ALL, however, will release all memory variables except those declared PUBLIC or initialized in a nested command file. >>> REPLACE REPLACE ALL does not replace all records correctly if an index is in use and the key field is replaced. Only the first record and those that logically follow the new value will be replaced. This occurs because the index is automatically updated (in-place key updating) when it is edited. The record pointer moves to the record following the new position, not to the record following the old position. This can be illustrated in the example given below (the data file has five records with the field CHARS-C-1, and is indexed on this field): . LIST Record# CHARS 1 a 2 b 3 c 4 i 5 j . REPLACE ALL Chars WITH 'd' 3 records replaced . LIST 2 b 3 c 1 d 4 d 5 d The manual warns against block replacements to the key field. The correct procedure would be to REPLACE with no indexes in use, open the indexes with SET INDEX TO, and then REINDEX. >>> REPORT FORM--Dates If you have a date-oriented report and you need to have it grouped by week, the following discussion will assist you. The grouping of dates into weeks has two requirements. First, the database file you are reporting from must be INDEXed on the date field that is being grouped on. Second, as the group expression in your REPORT FORM, you must have an expression that returns as its value the first day of the week for each date field. The expression is as follows: Yourdate - ( DOW( Yourdate ) - 1 ) When given any date value, this expression returns the date of the previous Sunday. It does this by subtracting from your date field the number of days that have passed since the last Sunday, the first day of the week in the dBASE III calendar. This value is obtained by subtracting 1 from the result of the DOW() function. If you wish to have the week you are grouping on start on a later day such as Monday, subtract more from the result of the DOW() function. For example, Monday would be DOW() - 2, Tuesday DOW() - 3, and so on. >>> REPORT FORM HEADING There appears to be some confusion about the use of the HEADING option to the REPORT FORM command. The argument of the HEADING statement is a character expression and, therefore, can contain any combination of memory variables, field variables including aliases, and functions that evaluate to a value of character type. This means that macro substitution is not necessary to in order to use variable data for the REPORT FORM in question. As a typical example: * ---Set up header string. title = "Sample Report Number 1" *---Run the report. REPORT FORM YourRpt HEADING title If the HEADING argument is a field variable from the database file being REPORTed on, the value of the HEADING will be the field value from the record being pointed to when the REPORT FORM is invoked. For example, if the current record number is 5 and a REPORT FORM is run with no scope, the HEADING value is set from record 5 and then the record pointer is reset to the top of file. The ability to use a database field as the HEADING argument presents some interesting possibilites. Suppose you have a master client and transaction file. You wish to have a transaction listing for a particular client with that client's name at the top of the REPORT FORM. The transactions database file is INDEXed in client number order. * ---Open database files. SELECT 1 USE Client SELECT 2 USE Transaction INDEX Transaction SELECT Client * ---Get the client to report on. LOCATE FOR Name = "Smith" * ---Report client's transactions. SELECT Transaction REPORT FORM Transaction HEADING Client->Name; WHILE Number = Client->Number >>> REPORT FORM, MODIFY REPORT (1) The report generator will right-justify field headings for numeric fields when the report is run. (2) If the PLAIN clause is specified with REPORT FORM TO PRINT, no page ejects occur. The report prints through to the end without page breaks. (3) MODIFY REPORT will allow the number of decimal places to be changed from the default. If this is done and the report is run, everything is as expected. However, if the report is modified again, the number of decimal places reverts to the default when the cursor reaches the "# decimal places" field. (4) Although not documented in the manual or in the cursor control menu, Ctrl-N inserts a column in a report being created or modified. However, its counterpart (Ctrl-U which deletes a column) is documented and included in the help menu. >>> REPORT FORM PLAIN The PLAIN option of the REPORT FORM command will cancel out the HEADING option when these two are used in the same command line. Therefore, do not use these two options in the same command. For example, the following command line will not print the HEADING "Week of May 6, 1985": REPORT FORM Wksales PLAIN HEADING "Week of May 6, 1985" TO PRINT >>> RANGE command The dBASE III Reference Manual states that RANGE may be used to specify lower and upper bounds for date variables, but does not clarify that the arguments of the RANGE clause involving literal dates must be in the form, CTOD("mm/dd/yy"). For example: @ 10,10 GET Mdate RANGE CTOD("01/01/85"), CTOD("01/01/86") >>> REPORT FORM The semicolon is not documented as functioning as a Carriage Return/Line-Feed in certain parts of REPORT FORMs. Menu 7 >>> Reserved Device Names in MS-DOS The MS-DOS manual specifies that certain names have a special meaning to MS-DOS. Do not use these reserved device names as dBASE III filenames: CON, AUX, COM1, COM2, LPT1, PRN, LPT2, LPT3, or NUL. This applies to database files, index files, command files, format files, or form files. Various error messages will result when files with any of these names are accessed. >>> Reserved words Page 1-138 of the tutorial in the first edition of the manual uses a sample routine which creates a memory variable with the name 'continue.' Since this is a reserved word, dBASE III will give the message, "No database in USE, enter filename." dBASE III is assuming you intend to CONTINUE on a LOCATE command. This will only happen if you use the = frm of assignment; dBASE III will execute correctly when you use the STORE TO form. Other words that will not work with the first syntax are: AVERAGE, COUNT, and SUM. >>> ROW(), COL() After a READ, the ROW() function always returns 24; however, the COL() function does not change. For example: SET TALK OFF var = SPACE(2) @ 5,40 GET var ? ROW(), COL() <--- This returns 6 and 3. READ ? ROW(), COL() <--- This returns 24 and 3. >>> RUN (or !) The RUN command requires that COMMAND.COM be in the boot drive or the directory indicated by SET COMSPEC. Otherwise, the incorrect error message "Insufficient memory" is displayed. >>> RUN COMMAND You can get the equivalent to Framework's DOS Access in dBASE III by issuing RUN COMMAND. This will leave you at the DOS operating system level and will allow you to enter any DOS commands. To get back to the dBASE III dot prompt, type EXIT. >>> SET ALTERNATE TO dBASE III will not send a linefeed (that is, CHR(10)) to an alternate file (WordPerfect looks for this linefeed character in its mail merge program). The following command file: SET ALTERNATE TO x SET ALTERNATE ON ?? "first LF" ?? CHR(10) ?? "second LF" SET ALTERNATE OFF CLOSE ALTERNATE will generate the following test file: first LFsecond LF As you can see, there is no linefeed in the file. >>> SET COLOR (1) To get enhanced video to be black on black, use the command SET COLOR TO x/y,0+/0. Black on black is frequently used to allow user input without displaying the characters on the screen, as with entry of a password. (2) The command: SET COLOR TO , will produce black on black, even if there is no space after the comma. (3) The asterisk (*) used with the SET COLOR command allows you to have blinking characters on the screen. The asterisk must be used in conjunction with a color parameter. For example: SET COLOR TO 6*/1,7/4,6 or: SET COLOR TO GR*/B,W/R,GR >>> SET COLOR TO on the Compaq computer SET COLOR TO U on the Compaq computer will not generate underlining. Instead, SET COLOR TO U and SET COLOR TO I give quarter intensity. Using U+ or I+ will generate half intensity. You will not get the monochrome attributes, because the Compaq monitor is operating as if it where color. >>> SET CONSOLE ON/OFF The SET CONSOLE ON/OFF command behaves differently in dBASE III than it does in dBASE II. Specifically, it has no effect when issued at the dot prompt, and if SET CONSOLE OFF is issued in a command file that neglects to SET CONSOLE ON, dBASE III will automatically set the console back on upon the termination of execution of that file. This includes normal termination as well as termination by means of pressing the escape key. >>> SET DEBUG ON SET DEBUG ON overrides SET CONSOLE OFF and output will be echoed to the screen. >>> SET DELETED SET DELETED in dBASE III applies to all commands, unlike dBASE II, where certain commands ignore the status of SET DELETED. In dBASE III you are permitted to COPY, APPEND and REPORT deleted records if DELETED is set OFF. >>> SET DOHISTORY ON/OFF (Developer's Release) If your dBASE III Developer's Release Reference Manual does not contain page 4-115A in the Commands section which explains the use of SET DOHISTORY, the following documents the command. SET DOHISTORY determines whether or not commands from command files are recorded in HISTORY. Syntax: SET DOHISTORY ON/OFF Defaults: SET DOHISTORY is normally OFF Usage: When SET DOHISTORY is ON, commands that were executed from within command files are stored in HISTORY. You can then edit these in full-screen edit mode and execute them. Note that SET DOHISTORY has no effect on commands you issue from the dot prompt. Tips: In conjunction with SUSPEND, SET DOHISTORY is useful for debugging. However, because SET DOHISTORY slows down command file execution, use it only when debugging. See Also: DISPLAY HISTORY, LIST HISTORY, RESUME, SET DEBUG, SET ECHO, SET HISTORY, SET STEP, SUSPEND >>> SET FILTER TO (1) The SORT command will only sort those records that meet the filter condition. The INDEX command will index all records in the file whether or not they meet the filter condition. However, a FIND command will return "No find" if the key falls outside the scope of FILTER. (2) Some users are confused regarding the performance that may be expected when using the SET FILTER TO command. The database file is not smaller in size when a filter is set. Consequently, activities that access the file are not proportionately faster. Commands that operate on the FILTERed set of records must still scan the entire database file. >>> SET MENUS ON/OFF The default for SET MENUS is OFF. However, ASSIST leaves MENUS SET ON even if they were off prior to entering ASSIST. >>> SET PROCEDURE TO, PARAMETERS, PROCEDURE The following program and procedure files illustrate the use of parameter passing with procedures. FUTVALUE.PRG calculates the future value of an investment and the future value of an annuity with the use of the BUSINESS.PRG procedure file. Notice how the parameters can be passed to BUSINESS.PRG. They can either be literal numbers, expressions or variables. Also, notice that the PARAMETERS command is included after each PROCEDURE that receives parameters. LISTINGS: * FUTVALUE.PRG * ------------ SET TALK OFF SET FIXED ON SET PROCEDURE TO Business * * { Calculate future value of an investment result = 0.00 DO FV WITH 6000.00, 8.5, 4, 5, result ? result * * { Calculate future value of regular deposits (Annuity) result = 0.00 DO FVA WITH 150.00, 7.0, 12, 2, result ? result * CLOSE PROCEDURE SET FIXED OFF SET TALK ON RETURN * EOF: FUTVALUE.PRG * BUSINESS.PRG { Library of business procedures * ------------ * PROCEDURE FV { Calculate future value of an investment PARAMETERS amount, rate, periods, years, result rate = rate / periods / 100 result = amount * (1 + rate) (periods * years) result = ROUND( result, 2 ) RETURN * PROCEDURE FVA { Calculate future value of regular deposits PARAMETERS amount, rate, periods, years, result rate = rate / periods / 100 result = amount * ( (1 + rate) (periods *; years) - 1 ) / rate result = ROUND( result, 2 ) RETURN * * EOF: BUSINESS.PRG OUTPUT: 9136.77 3852.15 Menu 8 >>> SET RELATION TO The SET RELATION TO command takes precedence over the SET FILTER TO and SET DELETED ON commands. That is, information from a related database file will be displayed if addressed with an alias name, even if that record has been filtered out with the SET FILTER TO command or with SET DELETED ON. Please note that because of this, SETting FILTER TO .NOT. DELETED() in another work area to exclude deleted records in that work area will not work if that record is accessed with the alias name. The following examples demonstrate that deleted and filtered records may be accessed if they are addressed with SET RELATION TO. Both database files consist of one 10-byte character field called One. SELECT 1 USE Test1 SELECT 2 USE Test2 INDEX ON One to Test2 DELETE ALL <-- Delete all from Test2. SET FILTER TO .NOT. DELETED() <-- Filter deleted records. SELECT 1 SET RELATION TO One INTO Test2 <-- SET RELATION. ? One, B->One AAAAAAAAAA AAAAAAAAAA ^--------------------- Field displays even though it is deleted and filtered from the Test2.DBF. >>> SET RELATION TO INTO (1) SET RELATION TO relates two database files by means of an expression. It can be used in conjunction with REPORT, LABEL, LIST, and many other commands to view information from more than one file. In the examples below, the command files on the left and right sides are identical in function. You will notice that it takes several more lines of code to accomplish the same task if the SET RELATION TO command is not used. Example 1: * ---Without SET RELATION. * ---Using SET RELATION. * ---Linked by RECNO(). * ---Linked by RECNO(). SELECT 2 SELECT 2 USE Address USE Address GO BOTTOM SELECT 1 STORE RECNO() TO last USE Names INDEX Alpha SELECT 1 SET RELATION TO RECNO() INTO Address USE Names INDEX Alpha DO WHILE .NOT. EOF() DO WHILE .NOT. EOF() ? Name,Address->Phone STORE RECNO() TO rec SKIP SELECT Address ENDDO IF rec <= last GO rec ELSE GO last SKIP ENDIF SELECT Names ? Name,Address->Phone SKIP ENDDO Example2: * ---Without SET RELATION * ---Using SET RELATION. * ---Linked on key field. * ---Linked on key field. SELECT 2 SELECT 2 USE Detail INDEX Custnbr USE Detail INDEX Alpha SELECT 1 SELECT 1 USE Customer USE Customer DO WHILE .NOT. EOF() SET RELATION TO Number INTO Detail STORE Number TO mkey DO WHILE .NOT. EOF() SELECT Detail ? Name,Detail->Phone SEEK mkey SKIP SELECT Customer ENDDO ? Name,Detail->Phone SKIP ENDDO To use SET RELATION TO with reports and labels, set the relation when the report or label is executed. When creating or modifying a report or label, only the structure of the active database file in the selected work area will be displayed at the top of the screen, so a listing of the unselected database file will be necessary. To specify fields from the unselected file, use the -> form. In order to print the report or label properly, both files will have to be in use and the relation will have to be set before the REPORT r LABEL command is issued. >>> SET UNIQUE ON/OFF The SET UNIQUE command determines whether duplicate key values will be stored in index files. The use of SET UNIQUE affects: 1. How the INDEX ON command constructs the named index file. 2. How the REINDEX and PACK commands rebuild all open index files in the currently selected work area. 3. How all open index files are maintained when the active database is updated (such as in APPEND, BROWSE, CHANGE, EDIT, or REPLACE). There is a tendency to believe that SET UNIQUE only needs to be used when creating an index file. SET UNIQUE does not work this way. The "uniqueness" of an index file is in no way attached to the file. When additions, deletions, or changes to key information are made to the database file, the index file must be open and SET UNIQUE must be ON or OFF (default). It is not possible to maintain index files of both unique and non-unique types for the same database file, unless one of them is recreated with SET UNIQUE ON or OFF every time a record is updated. Another item to consider when using "unique" and "non-unique" index files is when they are used in different SELECTed work areas. Updating indexed database files in different work areas will require that UNIQUE be SET to the appropriate value for each SELECTed area. >>> TRIM() with LIST or DISPLAY Due to the introduction of headings for LIST and DISPLAY, the TRIM() function may not work as you expect when used with these commands. Specifically, if you LIST or DISPLAY a list of TRIMmed fields, they will not be TRIMmed. Use an expression that includes the TRIM function instead. For example: LIST TRIM( Firstname ), Lastname will not work, but the following will: LIST TRIM( Firstname ) + ' ' + Lastname >>> VAL() function (1) In version 1.1, SET DECIMALS affects the VAL() function. Page 4-110 of the dBASE III User Manual states that SET DECIMALS applies only to division, SQRT(), LOG(), and EXP(). Page 5-39 says that the VAL() function does not display the decimal portion of a number. However, how the output of the VAL() is displayed will correspond to the number of decimal positions defined with the SET DECIMALS command. For example: SET DECIMALS TO x = "1.1111" ? VAL( x ) 1 SET DECIMALS TO 2 ? VAL( x ) 1.11 SET DECIMALS TO 10 ? VAL( x ) 1.1111000000 (2) Using any non-numeric argument with the VAL() function will return 0.00. For example: ? VAL("Truth") 0.00 However, if an asterisk is used as the argument of the VAL() function, it will return "****". ? VAL("*") **** In dBASE III, the asterisk is used in this way to signal that an operation resulted in a numeric overflow. If the VAL() function did not return asterisks, some operations might obscure the fact that a numeric overflow occurred. For example: divisor = 0 dividend = 1 x = dividend/divisor ? x ************** y = STR(x,1) ? y * z = VAL(y) ? z **** If the VAL() function above returned 0, the quotient might have been misinterpreted as being 0. (3) In converting from a character expression to a numeric value, the VAL() function evaluates the text from left to right until a non-numeric character is encountered. Leading blanks are ignored. If the function argument contains a mixture of number and text characters, only the leading numeric characters will be converted to a numeric value. For example: ? VAL( "12ABC" ) 12.00 Trailing blanks are treated as non-numeric characters and, when encountered, the conversion process is terminated. For example: ? VAL( "12 12" ) 12.00 >>> Warning against using a dBASE III file in dBASE II Whenever dBASE II attempts to read a database file, it writes a hexadecimal two (02H) to the first byte of the file. Therefore, attempting to USE a dBASE III file in dBASE II will make the file unusable by dBASE III. Any attempt to USE the corrupted file in dBASE III will give the error message "Not a dBASE database." Menu 9 INDEX OF dBASE III REFERENCE TOPICS Menu Choice Topic ------ -------------------- 1 @...GET...RANGE @...GET and READ @...GET...PICTURE @...SAY...PICTURE @...SAY using relative addressing APPEND FROM [SDF/DELIMITED [WITH ]] 2 Closing Database Files COPY TO [SDF/DELIMITED [WITH ]] COPY FILE TO COPY STRUCTURE EXTENDED / CREATE FROM CREATE / MODIFY REPORT Date conversion from dBASE II Dates that are blank Debugging tip Demonstration Disk (RunTime+) DISPLAY and LIST Duplicate Keywords in CONFIG.DB FILE() function 3 FIND / SEEK Function Keys Get Diskspace Get Current Directory Get Last Update Date and Time INPUT Macro (&) Substitution in a Format (.FMT) File MEMO fields MODIFY STRUCTURE 4 Numeric fields with decimal places Numeric input of large numbers PARAMETERS, passing Fields 5 PRIVATE PROCEDURE PROW() 6 PUBLIC Ramdisks RELEASE, incorrect syntax RELEASE ALL / CLEAR MEMORY REPLACE REPORT FORM--Dates REPORT FORM HEADING REPORT FORM / MODIFY REPORT REPORT FORM PLAIN 7 Reserved Device Names in MS-DOS Reserved words ROW() / COL() RUN (or !) RUN COMMAND SET ALTERNATE TO SET COLOR SET COLOR TO on the Compaq computer SET CONSOLE ON/OFF SET DEBUG ON SET DELETED SET DOHISTORY (Developer's Release) SET FILTER TO SET MENUS ON/OFF SET PROCEDURE TO / PARAMETERS, PROCEDURE 8 SET RELATION TO INTO SET RELATION TO SET UNIQUE ON/OFF TRIM with LIST or DISPLAY VAL() function Warning against using a dBASE III file in dBASE II 9 This Index. 10 CONFIG.SYS dCONVERT File Control Blocks Hard Disk Usage Installing version 1.0 Prolok (version 1.0) Menu 10 >>> CONFIG.SYS (1) The DOS configuration file, Config.sys, which contains FILES=20, must be in the root directory of the disk that DOS is either cold or warm booted from. The consequence of its absence when operating dBASE III is the unexpected error message "Too many files are open". To correct this problem: (a) Go to the root directory of the disk that DOS is booted from; (b) Create a Config.sys by entering the following at the system prompt: A>copy con: config.sys BUFFERS=15 FILES=20 Ctrl-Z (c) Reboot DOS by pressing Ctrl-Alt-Del. (2) Adding buffers and files with Config.sys is not free of memory overhead. Each buffer uses 528 bytes and each file uses approximately 39 bytes. In addition, hard disks such as Davong and Corvus use large amounts of memory. With BUFFERS and FILES set high, and a hard disk, some 256K systems may not have enough room to run dBASE III. Reducing these parameters to a lower amount (such as FILES=15 and BUFFERS=2) may be the only way to run dBASE III on these systems. >>> dCONVERT (1) dCONVERT, version 1.0 has an option for converting dBASE III files to dBASE II format. When this is chosen the first byte of the file is changed from 03 to 02 and other changes are made to the header, but a dBASE II format header is not created. The result is a file unusable by dBASE II or dBASE III. To work around this problem, COPY the dBASE III file to a text file using the SDF option. CREATE a file in dBASE II with the same structure and APPEND the text file using the SDF option. (2) dCONVERT will not convert command files that use four-letter abbreviations of dBASE II commands, such as DISP and ERAS. (3) Some 1-2-3 users have found that dCONVERT will not convert the dBASE II database files that 1-2-3's Translate utility generates. Apparently Translate creates a database file that is not an exact dBASE II database file. To work around this problem, create an ASCII text file from your 1-2-3 worksheet by using the following procedure: In Lotus 1-2-3: (1) Setup a worksheet to print without headers and footers with / P rint F ile O ther U nformatted; (2) Make the top, bottom, and left margins zero; (3) Write the text file with / P rint F ile; In dBASE III: (1) CREATE a database that reflects the structure of the 1-2-3 worksheet; (2) Read the text file into dBASE III with APPEND FROM .prn SDF. >>> FILE CONTROL BLOCKS dBASE III was designed to use the XENIX type "ASCII I/O handles" when setting up FCB's (File Control Blocks) for its I/O. Whenever dBASE III opens a file, DOS provides a control block. However, DOS only has space for 20 such blocks (FILES = 20). Since DOS uses five of these files, fifteen are left for dBASE III. >>> HARD DISK USAGE Customers using dBASE III with a hard disk system who wish to use an external editor for MODIFY COMMAND or memo field when editing will find that DOS checks the A: drive for COMMAND.COM whenever it loads the external editor if A: is the boot drive. This can be avoided by setting up the DOS environment to look for COMMAND.COM on the default drive (hard disk) and subdirectory. The command to do this is: SET COMSPEC=DRIVE\PATH\COMMAND.COM This can be entered at the system prompt or placed in an AUTOEXEC.BAT file. For example, if dBASE III were on drive C: in a subdirectory called DBASE, the command would be: SET COMSPEC=C:\DBASE\COMMAND.COM You must of course have COMMAND.COM present in this subdirectory. If the SET command is included in an AUTOEXEC.BAT file, "SET COMSPEC=" without any parameters must be included before the new command in order to clear the default setting. The status of this command can be checked by entering "SET" at the system prompt. >>> INSTALLING dBASE III (version 1.0) The errata sheet (dated June 18, 1984) that was shipped with the first several thousand copies of dBASE III explains a method for making the original system disk bootable using the SYS command. Unfortunately, this will not work. Since the directory space for the system was not reserved, the message, "No room for system on destination disk," results when SYS is run. In order to make bootable copies of dBASE III the following procedur must be followed: 1. Format two disks and place the operating system on them with the FORMAT d:/S command. 2. Copy DBASE.EXE and CONFIG.SYS onto one of the newly formatted disks. 3. Copy DBASE.OVL, CONFIG.DB, HELP.DBS, and ASSIST.HLP to the other new disk. 4. To boot dBASE III, place the copy of DBASE.EXE (disk from step 2. above) in drive A:, the original master system disk in drive B:, start the computer, and enter: DBASE The message "Insert dBASE III overlay file diskette, or type Ctrl-C to abort" will appear. 5. Remove the copy from drive A: and the master disk from drive B:. Place the disk with the copy of the overlay file (from step 3.) in drive A: and strike any key. 6. Place the master system disk in a safe place. It will not be needed until dBASE III is started again. Drive B: is free for whatever programs or data files you wish to work with. >>> PROLOK (version 1.0) (1) Each system disk that contains dBASE III is copy protected with the Prolok copy protection scheme. Every time dBASE III is loaded, Prolok which is internal to dBASE looks for the disk that has the Prolok identification on it before passing control to the dBASE loading routine. If the presence of the identifying disk is not sensed, the message "Unauthorized Duplicate" is displayed and the loading of dBASE is terminated. What must not be overlooked is that each copy of dBASE III is uniquely identified with the system disk that it came from, much as child to a parent. This is of particular importance to hard disk users who copy dBASE III to, and boot from the hard disk. Users booting from the hard disk will often attempt to use as the identifying disk a system disk that the dBASE on the hard disk cannot identify as its parent resulting in "Unauthorized Duplicate". (2) The Prolok procedure is designed to physically damage the disk. Therefore, running CHKDSK on it will report bad sectors -- 10240 bytes in bad sectors, to be exact.