joseLB 1 Report post Posted September 4, 2017 (edited) This UDf ( _FileDirectAccess.au3 ) implements direct access file method. It allows to create files with fixed-legth records with many fields, so you can write or read these records giving their record number. version 2 in sept 10 2017. I dis some importante changes. Records can have up to 180 fields. Each field can be: type 1= int32 (occupies 4 bytes in record) type 2= int64 (occupies 8 bytes in record) type 3= string (size defined by user) type 4= real - double (occupies 8 bytes) (not working yet, but for now you can use string with m chars -> ex: m=10, 16 or 20), where the most sgnificant digits are preserved (left digits) If there is any interest on them, I can put the formal UDF comments describing better the functions. How they work: open/create a DA file: _fileDA_Open (ByRef $file, $filename, $mode, $flush=10 ) Param1= variable previously created by user, ex: $file=0. It will be converted in an array, where all fle information is saved, including record fields. Param2= open mode = read (1) , read+write file (2), read+write+create (3) mode=1 => file must exists and can only be read, mode=2 => file + path will be created if does not exists. Record pointer will be after last record (so a write will be at file´s end), mode 3 => a previous file will be erased. File + path will be created. Param3= flush at each "n" writes. If 0, there are no flushes. default = 10 = a flush at each 10 writes define record fields and size: _FileDA_defineRecord(ByRef $file, $fields) Param1= array describing file, created by _fileDA_Open Param2= array [n][2]: n= # fields, [n][0]=field type, [n][1]= field lenght (used just if string, otherwise ignored) writes a record: _fileDA_Write (ByRef $file, ByRef $record) Param1= array describing file, created by _fileDA_Open Param2= array[n], n= number of record´s fields. [0]= field1, [1]= field2... each field must have same definition´s datatypes ** strings smaller than field´s size => will be filled with " ", bigger will be right trimmed (left n chars will remain) reads a record: _fileDA_Read (ByRef $file, ByRef $record) Param1= array describing file, created by _fileDA_Open Param2= array with # columns >= # fields. Each array´s column receives a record field Positions file pointer at next record to be read or write: _fileDA_Pos (ByRef $file, $pos, $origin=0) Param1= array describing file, created by _fileDA_Open Param2= pointer position. Param3= type of positioning: 0= absolute form start (default), 1= relative to the current position Closes file: _fileDA_Close (ByRef $file, $free=1) Param1= array describing file, created by _fileDA_Open Param2= type of close: 1 (defaut), clears $file array, so another file can be created on this variable. 0= another open is necessary, but defineRecord is not (ex: read) Besides attached UDF, there is na example that covers it´s usage. Spoiler expandcollapse popup#include <_FileDirectAccess.au3> Local $file1=0 ;must be declared. Wil be modified to an array by fileopen holding file definitions Local $file1_fields[4][2]=[ [3,10], [1,0], [2,0], [4,0] ] ;field definitions=> field1= string 10 chr, field2= int32, field3= int64, field4= double real (not working yet) Local $recwrite[4]= ["ABCDEFGHIJKLMN", 1023, 1024, 1024.5 ] ;record to write: 4 fields respecting field definitions (string can be bigger or smaller). Must be declared Local $recRead[4] ;array where record will be read. Must be declared. _ArrayDisplay($file1_fields, "fields definitions") ;show array with field definitions _ArrayDisplay($recwrite, "record to write") ;show record to write _fileDA_Open($file1 , "C:\GoogleDrive\bateria\VII\Central_PC\V6\Test1.txt" , 3) ;open a new file for read and write. Create file/path. If opened for read (1), file must exists _FileDA_defineRecord( $file1, $file1_fields ) ;defines each field type/size and record size = sum of fields sizes _fileDA_Write( $file1, $recwrite) ;writes 1st record $recwrite[0]= "second record" _fileDA_Write( $file1, $recwrite) ;writes 2nd record $recwrite[0]= "3th" _fileDA_Write( $file1, $recwrite) ;writes 3rd record (string smaller than 10 => will b filled with " " _fileDA_Pos($file1, 1,0) ;**** rewind pointer to 1st record to start read _fileDA_Read( $file1, $recRead ) ;reads 1st record _ArrayDisplay( $recRead, "record 1 readed") _fileDA_Read( $file1, $recRead ) ;reads 2nd record _ArrayDisplay( $recRead, "record 2 readed") _fileDA_Read( $file1, $recRead ) ;reads 3rd record _ArrayDisplay( $recRead, "record 3 readed") _fileDA_Pos($file1, 2 ,0) ;**** pointer set to re-read 2nd record _fileDA_Read( $file1, $recRead ) ;reads 2nd record _ArrayDisplay( $recRead, "rec. 2 re-readed") _fileDA_Close( $file1) ;+++++ speed tests +++++++ _fileDA_Open($file1 , "C:\GoogleDrive\bateria\VII\Central_PC\V6\Test2.txt" , 3) ;open a new file for read and write _FileDA_defineRecord( $file1, $file1_fields ) ;defines each field type/size and record size = sum of fields sizes $tempo= TimerInit() $recwrite[0]= "10K" For $K=1 to 10000 ;will record 10.000 records $recwrite[0]= $K/1.001 ;***string[n] can be a floating point (this example 10 lef most significant digits are saved, as string has 10 chars) $recwrite[1]= $K $recwrite[2]= $K+1 $recwrite[3]= $K/7 _fileDA_Write( $file1, $recwrite) Next MsgBox(4096,"10K rec. writes", "time to write (ms)= " & int(TimerDiff( $tempo))) ; time spent recording $tempo= TimerInit() _fileDA_Pos($file1, 1 ,0) ;rewind to record 1 to read 10K records and measure their time For $K= 1 To 10000 ; time to read 10.000 records _fileDA_Read( $file1, $recRead ) Next MsgBox(4096,"10K rec. reads", "time to read (ms)= " & int(TimerDiff( $tempo))) For $K=1 To 7 $R= Random(1,10000,1) _fileDA_Pos($file1, $R ,0) ;randomic record _fileDA_Read( $file1, $recRead ) _ArrayDisplay( $recRead, "rec." & $R & " OK?") Next MsgBox(4096, "skipped writes", "up to now 10K regs where writted sequentialy" &@cr& "now just reg 15K and 14K would be writted" &@cr& "and some reads will be done") _fileDA_Pos($file1, 15000 ,0) ;write a new record "skipping" 5K records (rec. 15K) $recwrite[1]= 15000 _fileDA_Write( $file1, $recwrite) _fileDA_Pos($file1, 14000 ,0) ;write a new record (rec. 14K) $recwrite[1]= 14000 _fileDA_Write( $file1, $recwrite) _fileDA_Pos($file1, 15000 ,0) _fileDA_Read( $file1, $recRead ) ;reads rec. 15K to see if OK _ArrayDisplay( $recRead, "rec 15K" ) _fileDA_Pos($file1, 13000 ,0) ;reads non-writed record 13K _fileDA_Read( $file1, $recRead ) _ArrayDisplay( $recRead, "rec 13K not writed" ) _fileDA_Pos($file1, 14000 ,0) ;reads record 14K _fileDA_Read( $file1, $recRead ) _ArrayDisplay( $recRead, "rec 14K" ) _fileDA_Close( $file1) version 2 : _FileDirectAccess.au3 Edited September 11, 2017 by joseLB Share this post Link to post Share on other sites
joseLB 1 Report post Posted September 10, 2017 (edited) About real/float/double numbers that are not working: ** meanwile you can use a string field type with the desired size to preserve the precision you want: ex: 1024.123456 will need 11 chars The real (float) seems to be writen OK. So, the problem is to read them. In a record, where the real/double number writen was 1024.5 using FileWrite( $file[1] , Number($record[$K], 3 ) ) , later reading it with $X= FileRead( $file[1], 8) , and observing $X in MsgBox(...) = 0x0000000000029040 that seems to be a real number, as: 1024.5 = 0x0...29040 1025.5 = 0x0...69040 1026.5 = 0x0...A9040 I suppose I tried all conversion routines avaiable... any clue? I googled for some float point converters, but could reach a resonable answer. Thanks in advance. Edited September 10, 2017 by joseLB Share this post Link to post Share on other sites
