MT5 RAM memory voraciousness, problems with reading/writing large files
At least put the code in MQL... You have to count line by line.
SQL, Apache Spark or other variants are used for such tasks.
At least put the code in MQL ... Line by line should be calculated.
For such tasks and uses SQL, Apache Spark or other options.
I don't have calculations there yet, it's just that writing to the array takes so much time for unknown reasons (actually, it's reading from the file that eats up the time, it's not clear about memory yet).
I don't rule out a problem with the code. I am using read class, tweaked (kinda rules) from MT4.
The main code is almost 700 lines - nothing interesting there - we declare array, reset array to zero, change array size, write data to array from file, overwrite data from array to array, write data from array to file.
Here is the class that was paid for:
//+------------------------------------------------------------------+ //| CSVReader.mqh | //| Copyright 2015, MetaQuotes Software Corp. | //| http://www.mql5.com | //+------------------------------------------------------------------+ #property copyright "Copyright 2015, MetaQuotes Software Corp." #property link "http://www.mql5.com" #property strict //+------------------------------------------------------------------+ //| CSVReader | //+------------------------------------------------------------------+ class CSVReader { private: int m_flags; string m_filename; char m_delimiter; bool m_data_loaded; //--- string m_column_names[]; string m_cells[]; //--- int m_total_rows; int m_total_columns; //--- bool AddData(string data_str,bool header); bool CheckIndexes(int row,int column); string ReplaceCommaByDot(string str); public: CSVReader(string path); //--- void SetDelimiter(char delimiter); void SetCommon(bool common); //--- bool Load(int start_line); //--- string GetStringValue(int row,int column); int GetIntValue(int row,int column); double GetDoubleValue(int row,int column); bool GetBoolValue(int row,int column); //--- int TotalRows() {return(m_total_rows);} int TotalColumns() {return(m_total_columns);} //--- string FileName() {return(m_filename);} //--- string ColumnName(int index); }; //+------------------------------------------------------------------+ //| CSVReader class constructor | //+------------------------------------------------------------------+ void CSVReader::CSVReader(string filename) { m_filename=filename; m_flags=FILE_SHARE_READ|FILE_TXT|FILE_ANSI; m_delimiter=';'; m_data_loaded=false; ArrayResize(m_column_names,0); ArrayResize(m_cells,0); m_total_rows=0; m_total_columns=0; } //+------------------------------------------------------------------+ //| SetDelimiter | //+------------------------------------------------------------------+ void CSVReader::SetDelimiter(char delimiter) { m_delimiter=delimiter; } //+------------------------------------------------------------------+ //| SetCommon | //+------------------------------------------------------------------+ void CSVReader::SetCommon(bool common) { if(common) m_flags|=FILE_COMMON; else m_flags&=~FILE_COMMON; } //+------------------------------------------------------------------+ //| ColumnName | //+------------------------------------------------------------------+ string CSVReader::ColumnName(int index) { if(m_total_columns==0) return(""); if((index>=0) && (index<m_total_columns)) return(m_column_names[index]); return("error"); } //+------------------------------------------------------------------+ //| AddData | //+------------------------------------------------------------------+ bool CSVReader::AddData(string data_str,bool header) { string str=data_str; int DlinaStroki=StringLen(str); int ObrezLeft=StringTrimLeft(str); int ObrezRifgr=StringTrimRight(str); if(ObrezLeft>0) { str=StringSubstr( str, // строка ObrezLeft, // с какой позиции начать DlinaStroki-ObrezLeft // длина извлекаемой строки ); ObrezRifgr=StringTrimRight(str); } if(ObrezRifgr>0) { str=StringSubstr( str, // строка 0, // с какой позиции начать DlinaStroki-ObrezRifgr // длина извлекаемой строки ); } // str=StringTrimLeft(str); // str=StringTrimRight(str); //--- if(StringLen(str)==0) return(false); string lines[]; StringSplit(str,m_delimiter,lines); //--- check ending with delimiter if(StringLen(str)>0) { string str0=StringSubstr(str,StringLen(str)-1,1); //--- ending with delimiter case, decrease lines[] if(str0==CharToString(m_delimiter)) { ArrayResize(lines,ArraySize(lines)-1); } } //--- parse header if(header==true) { m_total_columns=ArraySize(lines); if(m_total_columns==0) return(false); ArrayResize(m_column_names,m_total_columns); //--- for(int i=0; i<m_total_columns; i++) { m_column_names[i]=lines[i]; } } else //--- parse data { int columns_count=ArraySize(lines); //--- add data if(columns_count==m_total_columns) { m_total_rows++; ArrayResize(m_cells,m_total_rows*m_total_columns,10000); //--- for(int i=0; i<columns_count; i++) { int index=m_total_columns*(m_total_rows-1)+i; if((index>=0) && (index<ArraySize(m_cells))) { m_cells[index]=ReplaceCommaByDot(lines[i]); } else return(false); } } } //--- return(true); } //+------------------------------------------------------------------+ //| Load | //+------------------------------------------------------------------+ bool CSVReader::Load(int start_line) { int filehandle=FileOpen(m_filename,FILE_CSV|FILE_READ|FILE_ANSI|FILE_SHARE_READ,'\n'); if(filehandle==INVALID_HANDLE) { Alert("Error in open of file ",m_filename,", error",GetLastError()); return(false); } //--- int line_index=0; while(!FileIsEnding(filehandle)) { string str=FileReadString(filehandle); //--- skip 0th row if(line_index>=start_line) if(str!="") { if(line_index==1) AddData(str,true); else AddData(str,false); } line_index++; } //--- FileClose(filehandle); return(true); } //+------------------------------------------------------------------+ //| ReplaceCommaByDot | //+------------------------------------------------------------------+ string CSVReader::ReplaceCommaByDot(string str) { string str0=""; for(int i=0; i<StringLen(str); i++) { ushort chr=StringGetCharacter(str,i); if(chr==',') chr='.'; str0+=CharToString((uchar)chr); } return(str0); } //+------------------------------------------------------------------+ //| CheckIndexes | //+------------------------------------------------------------------+ bool CSVReader::CheckIndexes(int row,int column) { if((m_total_columns==0) || (m_total_rows==0)) return(false); if((row<0) || (row>=m_total_rows)) return(false); if((column<0) || (column>=m_total_columns)) return(false); //--- return(true); } //+------------------------------------------------------------------+ //| GetStringValue | //+------------------------------------------------------------------+ string CSVReader::GetStringValue(int row,int column) { if(CheckIndexes(row,column)==false) return(""); //--- int index=m_total_columns*row+column; if((index>=0) && (index<ArraySize(m_cells))) return(m_cells[index]); return("error"); } //+------------------------------------------------------------------+ //| GetIntValue | //+------------------------------------------------------------------+ int CSVReader::GetIntValue(int row,int column) { if(CheckIndexes(row,column)==false) return(0); //--- int index=m_total_columns*row+column; if((index>=0) && (index<ArraySize(m_cells))) { return((int)StringToInteger(m_cells[index])); } return(0); } //+------------------------------------------------------------------+ //| GetDoubleValue | //+------------------------------------------------------------------+ double CSVReader::GetDoubleValue(int row,int column) { if(CheckIndexes(row,column)==false) return(0.0); //--- int index=m_total_columns*row+column; if((index>=0) && (index<ArraySize(m_cells))) { return(StringToDouble(m_cells[index])); } return(0.0); } //+------------------------------------------------------------------+ //+------------------------------------------------------------------+ //| GetBoolValue | //+------------------------------------------------------------------+ bool CSVReader::GetBoolValue(int row,int column) { if(CheckIndexes(row,column)==false) return(0.0); //--- int index=m_total_columns*row+column; if((index>=0) && (index<ArraySize(m_cells))) { if (StringToInteger(m_cells[index])==1) return (true); else return (false); } return(0.0); } //+------------------------------------------------------------------+
Script (writing is omitted, because the problem is with reading) is a stripped-down version.
#property version "1.00" #property indicator_chart_window #property strict //--- #property script_show_inputs #include <CSVReader.mqh> //Класс по чтению информации из файла //#include <CsvWriter_2.mqh> //Класс по записи информации в файл input string FileLoadSetup="PredFind\\Test\\Pred_Ocenka_Read.csv"; input int Prediktor=1; //Целевая input double Support=0.01;//Поддержка input double Relevantnost=70.0;//Достоверность int arrRead_01[]; int arrRead_02[]; int arrRead_03[]; //--Информационные string arrRead_DateTime[]; int arrWrite_01[]; int arrWrite_02[]; int arrWrite_03[]; //--Информационные string arrWrite_DateTime[]; CsvWriter Printer; int Statistic; int StrokTotal=0; int arrSize=0; //+------------------------------------------------------------------+ //| Script program start function | //+------------------------------------------------------------------+ void OnStart() { CSVReader FileLoads(FileLoadSetup); FileLoads.SetDelimiter(';'); FileLoads.SetCommon(true); int StrokaSetup=0; if(FileLoads.Load(1)) { StrokTotal=FileLoads.TotalRows(); arrSize=StrokTotal; PrintFormat("File%s loaded. Total rows=%d,Total columns=%d",FileLoads.FileName(),FileLoads.TotalRows(),FileLoads.TotalColumns()); ArrayFree(arrRead_01); ArrayFree(arrRead_02); ArrayFree(arrRead_03); ArrayFree(arrRead_DateTime); ArrayFree(arrWrite_01); ArrayFree(arrWrite_02); ArrayFree(arrWrite_03); ArrayFree(arrWrite_DateTime); //--Зададим размер массивам ArrayResize(arrRead_01,arrSize); ArrayResize(arrRead_02,arrSize); ArrayResize(arrRead_03,arrSize); ArrayResize(arrRead_DateTime,arrSize); ArrayResize(arrWrite_01,arrSize); ArrayResize(arrWrite_02,arrSize); ArrayResize(arrWrite_03,arrSize); ArrayResize(arrWrite_DateTime,arrSize); for(int i=1;i<StrokTotal; i++) { //Print(FileLoads.GetIntValue(i,20)); } for(int i=1;i<StrokTotal; i++) { arrRead_DateTime[i]=FileLoads.GetStringValue(i,0); arrRead_01[i]=FileLoads.GetIntValue(i,1); arrRead_02[i]=FileLoads.GetIntValue(i,2); arrRead_03[i]=FileLoads.GetIntValue(i,3); } } for(int i=1;i<arrSize; i++) { arrWrite_DateTime[i]=arrRead_DateTime[i]; arrWrite_01[i]=arrRead_01[i]; arrWrite_02[i]=arrRead_02[i]; arrWrite_03[i]=arrRead_03[i]; } } //+------------------------------------------------------------------+
There is one more problem - I can't figure out how to write more than 48 columns. I also use the class. Can anyone help correct the class in order to expand the number of columns for the record?
Here is the code, which is missing in the previous post, to write to a .
string TimeF= TimeToString ( TimeLocal (), TIME_DATE | TIME_MINUTES ); StringSetCharacter (TimeF, 13 , '_' ); Statistic=Printer.FileCreate( Symbol ()+ "_" + "Pred_Ocenka_Write" , "PredFind\\Test" , false , false ,EvryTick); //Создание файла для записи Printer.Write( "arr_Buy_Sell" , "arr_Vektor_Week" , "arr_Vektor_Day" , "arr_Vektor_Don" , "arr_DonProc" , "arr_iDelta_H_1_H1" , "arr_iDelta_H_1_H3" , "arr_iDelta_H_1_H4" , "arr_iDelta_H_1_H6" , "arr_iDelta_H_1_H12" , "arr_iDelta_H_1_D1" , "arr_iDelta_H_1_W1" , "arr_iDelta_H_1_MN1" , "arr_RSI_Open_M1" , "arr_RSI_Open_H1" , "arr_BB_Center" , "arr_BB_Up" , "arr_BB_Down" , "arr_TimeH" , "arr_Den_Nedeli" , "arr_iDelta_Max_H1" , "arr_iDelta_Min_H1" , "arr_iDelta_Max_D1" , "arr_iDelta_Min_D1" , "arr_Buy Расчетная" , "arr_Buy Номер правила" , "arr_Buy Поддержка" , "arr_Buy Достоверность" ); for ( int i= 1 ;i<arrSize; i++) { Printer.Write( arrRead_DateTime[i], IntegerToString (arrWrite_01[i]), IntegerToString (arrWrite_02[i]), IntegerToString (arrWrite_03[i]), IntegerToString (arrWrite_04[i]), IntegerToString (arrWrite_05[i]), IntegerToString (arrWrite_06[i]), IntegerToString (arrWrite_07[i]), IntegerToString (arrWrite_08[i]), IntegerToString (arrWrite_09[i]), IntegerToString (arrWrite_10[i]), IntegerToString (arrWrite_11[i]), IntegerToString (arrWrite_12[i]), IntegerToString (arrWrite_13[i]), IntegerToString (arrWrite_14[i]), IntegerToString (arrWrite_15[i]), IntegerToString (arrWrite_16[i]), IntegerToString (arrWrite_17[i]), IntegerToString (arrWrite_18[i]), IntegerToString (arrWrite_19[i]), IntegerToString (arrWrite_20[i]), IntegerToString (arrWrite_21[i]), IntegerToString (arrWrite_22[i]), IntegerToString (arrWrite_23[i]), IntegerToString (arrWrite_24[i]), IntegerToString (arrWrite_25[i]), IntegerToString (arrWrite_26[i]), IntegerToString (arrWrite_27[i]), IntegerToString (arrWrite_28[i]), IntegerToString (arrWrite_29[i]), IntegerToString (arrWrite_30[i]), IntegerToString (arrWrite_40[i]), IntegerToString (arrWrite_41[i]), IntegerToString (arrWrite_42[i]), IntegerToString (arrWrite_43[i]), IntegerToString (arrWrite_44[i]), IntegerToString (arrWrite_45[i]), IntegerToString (arrWrite_46[i]), IntegerToString (arrWrite_47[i]), IntegerToString (arrWrite_48[i]), IntegerToString (arrWrite_49[i]), IntegerToString (arrWrite_50[i]), IntegerToString (arrWrite_51[i]), IntegerToString (arrWrite_52[i]), IntegerToString (arrWrite_53[i]), IntegerToString (arrWrite_54[i]), IntegerToString (arrWrite_55[i]), IntegerToString (arrWrite_56[i]) ); }
But the class for writing data to a file is also paid
//+------------------------------------------------------------------+ //| CsvWriter.mqh | //| Виктор Крупинский | //| Krupinskiy@Mail.Ru | //+------------------------------------------------------------------+ #property copyright "Виктор Крупинский" #property link "Krupinskiy@Mail.Ru" #property version "2.00" #property strict enum periodicity { EvryTick=- 1 ,NewBar= PERIOD_CURRENT ,EvryHour= PERIOD_H1 , EvryDay= PERIOD_D1 ,EvryWeek= PERIOD_W1 ,EvryMonth= PERIOD_MN1 }; class CsvWriter { private : int handle; periodicity writeperiod; datetime opentime; bool flag; string DotToComma( string value); public : CsvWriter() { flag= false ; }; ~CsvWriter() { if (flag) FileClose (handle); }; int FileCreate( string fileame, string Dir= "" , bool Common= false , bool FileNew= true ,periodicity period=EvryTick); void FileSelectByHandle( int h,periodicity period=EvryTick) { handle=h; SetPeriodicity(period);}; void SetPeriodicity(periodicity period); int Write( string T1, string T2= "" , string T3= "" , string T4= "" , string T5= "" , string T6= "" , string T7= "" , string T8= "" , string T9= "" , string T10= "" , string T11= "" , string T12= "" , string T13= "" , string T14= "" , string T15= "" , string T16= "" , string T17= "" , string T18= "" , string T19= "" , string T20= "" , string T21= "" , string T22= "" , string T23= "" , string T24= "" , string T25= "" , string T26= "" , string T27= "" , string T28= "" , string T29= "" , string T30= "" , string T31= "" , string T32= "" , string T33= "" , string T34= "" , string T35= "" , string T36= "" , string T37= "" , string T38= "" , string T39= "" , string T40= "" , string T41= "" , string T42= "" , string T43= "" , string T44= "" , string T45= "" , string T46= "" , string T47= "" , string T48= "" , string T49= "" , string T50= "" , string T51= "" , string T52= "" , string T53= "" , string T54= "" , string T55= "" , string T56= "" , string T57= "" , string T58= "" , string T59= "" , string T60= "" , string T61= "" , string T62= "" , string T63= "" ) { int ReturnValue= 0 ; if (writeperiod==- 1 || opentime!=iTime_Old( _Symbol ,writeperiod, 0 )) { if ( FileWrite (handle,DotToComma(T1),DotToComma(T2),DotToComma(T3),DotToComma(T4),DotToComma(T5),DotToComma(T6), DotToComma(T7),DotToComma(T8),DotToComma(T9),DotToComma(T10),DotToComma(T11),DotToComma(T12), DotToComma(T13),DotToComma(T14),DotToComma(T15),DotToComma(T16),DotToComma(T17),DotToComma(T18), DotToComma(T19),DotToComma(T20),DotToComma(T21),DotToComma(T22),DotToComma(T23),DotToComma(T24), DotToComma(T25),DotToComma(T26),DotToComma(T27),DotToComma(T28),DotToComma(T29),DotToComma(T30), DotToComma(T31),DotToComma(T32),DotToComma(T33),DotToComma(T34),DotToComma(T35),DotToComma(T36), DotToComma(T37),DotToComma(T38),DotToComma(T39),DotToComma(T40),DotToComma(T41),DotToComma(T42), DotToComma(T43),DotToComma(T44),DotToComma(T45),DotToComma(T46),DotToComma(T47),DotToComma(T48), DotToComma(T49),DotToComma(T50),DotToComma(T51),DotToComma(T52),DotToComma(T53),DotToComma(T54), DotToComma(T55),DotToComma(T56),DotToComma(T57),DotToComma(T58),DotToComma(T59),DotToComma(T60), DotToComma(T61),DotToComma(T62),DotToComma(T63))) { ReturnValue= 1 ; FileFlush (handle); } } opentime = writeperiod==EvryTick ? 0 : iTime_Old( _Symbol ,writeperiod, 0 ); return (ReturnValue); }; //--------------------------------- End Write ------------------------------------- }; //--------------------- End CsvWriter ------------------------------------------------ int CsvWriter::FileCreate( string filename, string Dir= "" , bool Common= false , bool FileNew= true , periodicity period=EvryTick) { //Print ("Попытка создать файл"); if (FileNew) { //Print ("Новый файл?"); string NextName; int Pos; string Filter = Dir== "" ? "*" : Dir+ "\\*" ; long SearchHandle=Common ? FileFindFirst (Filter,NextName, FILE_COMMON ) : FileFindFirst (Filter,NextName); int Count = StringFind (NextName,filename)>= 0 && (Pos= StringFind (NextName, "_" ))> 0 ? ( int ) StringSubstr (NextName, 0 ,Pos) : 0 ; NextName= "" ; while ( FileFindNext (SearchHandle,NextName)) if ( StringFind (NextName,filename)>= 0 && (Pos= StringFind (NextName, "_" ))> 0 ) Count= MathMax (Count,( int ) StringSubstr (NextName, 0 ,Pos)); FileFindClose (SearchHandle); Count++; filename= IntegerToString (Count, 6 , '0' )+ "_" +filename; } filename = (Dir== "" ? filename : Dir+ "\\" +filename)+ ".csv" ; handle= Common ? FileOpen (filename, FILE_CSV | FILE_COMMON | FILE_READ | FILE_WRITE | FILE_SHARE_READ ) : FileOpen (filename, FILE_CSV | FILE_READ | FILE_WRITE | FILE_SHARE_READ ); // Print ("handle=",handle); flag=handle> 0 ; if (handle> 0 ) { FileSeek (handle, 0 , SEEK_END ); ulong pos= FileTell (handle); SetPeriodicity(period); } return (handle); } //---------------------- End FileCreate ------------------------------------------------ void CsvWriter::SetPeriodicity(periodicity period) { writeperiod = period; opentime = period==EvryTick ? 0 : iTime_Old( _Symbol ,period, 0 ); } //---------------------- End SetPeriodicity -------------------------------------------- string CsvWriter::DotToComma( string value) { int Pos= 0 ; if ((Pos= StringFind (( string )value, "." ))>= 0 && StringFind ( StringSubstr (( string )value,Pos+ 1 ), "." )< 0 ) //value=StringSetChar(value,Pos,','); //Print (value); StringSetCharacter (value,Pos, ',' ); //Print (value); return (value); } //---------------------- End DotToComma ------------------------------------------------ //+------------------------------------------------------------------+ //| | //+------------------------------------------------------------------+ datetime iTime_Old( string symbol, int tf, int index) { if (index < 0 ) return (- 1 ); ENUM_TIMEFRAMES timeframe=TFMigrate_Old(tf); datetime Arr[]; if ( CopyTime (symbol,timeframe,index, 1 ,Arr)> 0 ) return (Arr[ 0 ]); else return (- 1 ); } //+------------------------------------------------------------------+ //| | //+------------------------------------------------------------------+ ENUM_TIMEFRAMES TFMigrate_Old( int tf) { switch (tf) { case 0 : return ( PERIOD_CURRENT ); case 1 : return ( PERIOD_M1 ); case 5 : return ( PERIOD_M5 ); case 15 : return ( PERIOD_M15 ); case 30 : return ( PERIOD_M30 ); case 60 : return ( PERIOD_H1 ); case 240 : return ( PERIOD_H4 ); case 1440 : return ( PERIOD_D1 ); case 10080 : return ( PERIOD_W1 ); case 43200 : return ( PERIOD_MN1 ); case 2 : return ( PERIOD_M2 ); case 3 : return ( PERIOD_M3 ); case 4 : return ( PERIOD_M4 ); case 6 : return ( PERIOD_M6 ); case 10 : return ( PERIOD_M10 ); case 12 : return ( PERIOD_M12 ); case 16385 : return ( PERIOD_H1 ); case 16386 : return ( PERIOD_H2 ); case 16387 : return ( PERIOD_H3 ); case 16388 : return ( PERIOD_H4 ); case 16390 : return ( PERIOD_H6 ); case 16392 : return ( PERIOD_H8 ); case 16396 : return ( PERIOD_H12 ); case 16408 : return ( PERIOD_D1 ); case 32769 : return ( PERIOD_W1 ); case 49153 : return ( PERIOD_MN1 ); default : return ( PERIOD_CURRENT ); } }I have attached the whole script.
78 Mbytes - there must be about 1,000,000 lines.
First the whole file is copied into memory. Then 4 arrays of int type (8 bytes) multiplied by the number of lines. Of course, the RAM will run out quickly.
To work with large files, reading and writing should be done sector by sector, and here even unnecessary memory is not freed up.78 Mbytes - there must be about 1,000,000 lines in there.
First the whole file is copied into memory. Then 4 arrays of int type (8 bytes) multiplied by the number of lines. Of course, the RAM will run out quickly.
To work with large files I need to read and write sector by sector, but here even unnecessary memory is not freed up.There are 606004 lines in the file.
Doesn't exlle copy data into memory?
Maybe we'd have to handle the file in some other way, since the most time is spent reading data from the file and I don't understand why it's so.
I need complete array in memory for further work, otherwise I spend much more time for reading/writing.
I have a question, my CSV file is 78 megabytes in size, I want to calculate it, as Excel is very slow because of its size, I was thinking to use MT5.
I made a script that only reads data into arrays for each line (57 arrays for reading from file and 57 for writing into new file), the size of array is equal to the number of lines in CSV file. So, the script ate 8 Gbytes of memory, while Excel was less than 1.5 Gbytes, and the read/write operation took MT5 as much as 9 minutes!
You have 600 000 strings, each containing up to 57 cells which are read into a one-dimensional array string m_cells[]; That is 36 million string elements in the array. The string type , I recall, takes 12 bytes even for an empty string. That's 400 MB of service (addresses, string lengths) information alone, even if all cells are empty. See what happens when you do a resize of this array when you add one new row:
m_total_rows++;
ArrayResize(m_cells,m_total_rows*m_total_columns,10000);
- for each of the 36 million row-cells, you have to find a chunk of free memory long enough, grab it, copy the value to a new location, and spell the length. The fact that 10 thousand elements (or bytes, I do not remember) more memory is allocated doesn't matter: when one more row comes, m_total_rows will increase again, m_total_rows*m_total_columns too, and again memory for the array will be reallocated by copying. This is the reason why the script takes so long.
Memory is allocated in chunks multiple of 8 bytes even for 32-bit operating systems. We can assume that m_cells is not (400 + 78) MB, but 1 Gb. For each copy, we need as much more memory space as m_cells already occupies, doubling the memory would take up 2 GB of memory. Instead of 78 MB, which contains the actual information.
All in all, it's a very inefficient way to store in string cells. It is difficult to be more specific without knowing the task.
Try to eliminate functions ArraySize() and ArrayResize() - you know the size. To be on the safe side, take 2 times as much. Report results
I have a question, my CSV file is 78 megabytes in size, I want to calculate it, as Excel is very slow because of its size, I was thinking to use MT5.
I made a script that only reads data into arrays for each line (57 arrays for reading from file and 57 for writing into new file), the size of array is equal to the number of lines in CSV file. So, the script ate 8 Gbytes of memory, while Excel was less than 1.5 Gbytes, and the read/write operation took MT5 as much as 9 minutes!
I had this problem once. Did a lot of fixes. Can't remember the whole thing - it was a one-off job,
but try to setArrayResize(arrRead_01,arrSize); 3 parameter
|
Experiment.
You have 600 thousand strings, each containing up to 57 cells, which are read into a one-dimensional array string m_cells[]; That is, 36 million string elements in the array. The string type , I recall, takes 12 bytes even for an empty string. That's 400 MB of service (addresses, string lengths) information alone, even if all cells are empty. See what happens when you do a resize of this array when you add one new row:
m_total_rows++;
ArrayResize(m_cells,m_total_rows*m_total_columns,10000);
- for each of the 36 million row-cells, you have to find a chunk of free memory long enough, grab it, copy the value to a new location, and spell the length. The fact that 10 thousand elements (or bytes, I do not remember) more memory is allocated doesn't matter: when one more row comes, m_total_rows will increase again, m_total_rows*m_total_columns too, and again memory for the array will be reallocated by copying. This is the reason for the script's slowness.
Thanks for the code analysis.
I'm having a hard time understanding OOP, so making edits to it is torture for me. As I understood it's about the class, which is responsible for reading, that's where the whole file is split into string arrays, and only then, if I want, I convert these string values to another type and put them into arrays of int type in my script. If I understand correctly, then to speed up the process, I should read the entire file and figure out how many rows and columns are there and how many symbols are in each cell, and then immediately select the array size? Then how can we do it?
Memory is allocated in chunks multiple of 8 bytes even for 32-bit operating systems. We may suppose that m_cells will occupy 1 Gb, not (400 + 78) Mb. For each copy, we need as much more memory space as m_cells already occupies, doubling the memory would take up 2 GB of memory. Instead of the 78 MB that contains the actual information.
All in all, a very inefficient way to store in string cells. It is difficult to be more specific without knowing the task.
I'd like to understand where 400 megabytes is coming from all at once? Did I understand correctly that memory is allocated, but not cleared/released previously allocated memory?
Try to throw out ArraySize() and ArrayResize() - you know the size. Take 2 times bigger just in case... Let me know the result
From where to dump from a script or class? The size is known, since I know the file size? But, the file may be a different size, the script is not done at once...
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
You agree to website policy and terms of use
I have a question, my CSV file is 78 megabytes in size, I want to calculate it, as Excel is very slow because of its size, I was thinking to use MT5.
I made a script that only reads data into arrays for each line (57 arrays for reading from file and 57 for writing into new file), the size of array is equal to the number of lines in CSV file. So, the script ate 8 gigabytes of memory, while Excel is less than 1.5 gigabytes, and the read/write operation took MT5 as much as 9 minutes!