asked on

ignore comments in a text file

I have a standard text file containing integers.
The file has no consitant structure, and I would like to put comments in it, preceded by a semicolon.

for example:

2 ;input units
50 150
50 400
3 ;process units
200 300
200 375
200 450

I would like to read the integers into an array, but ignore the comments.

Any suggestions appreciated.

JonathanH13

ASKER CERTIFIED SOLUTION

Russell Libby

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial

TheRealLoki

I think this should do everything you need

var
i, j: integer;
s, s2: string;
sl: tstringlist;
begin
sl := tstringlist.create;
try
sl.loadfromfile('somefile.txt');
for i := 0 to pred(sl.count) do
begin
s := trim(sl[i]); // take 1 line
j := pos(';', s);
if j <> 0 then
begin // remove comment
s := trim(copy(s, 1, j-1));
end;
j := pos(' ', s); // start check for space
while (j > 0) do
begin
s2 := copy(s, 1, j-1); // coy this number to s2 so we can work with it
delete(s, 1, j); // remove this number and the space after it
memo1.lines.add(strToInt(s2)); // display this number in our memo - this is just as an example
j := pos(' ', s); // check for another space. ie. another number follows
end;
if s <> '' then memo1.lines.add(strtoint(s)); // the last number - dont forget to use this
end;
finally
sl.clear;
sl.free;
end;

Tueblo

This should do the job:

TDataArray = array of Integer;

function IsWhitespace(c: Char): Boolean;
begin
result := (c = #09) or (c = #32); // add more whitespaces if neccessary
end;

function GetAndDelNextLine(var str: String): String;
const cEOL = #13#10; // End Of Line sign
var p: Integer;
begin
result := '';
p := Pos(cEOL, str);
if (p > 0) then
begin
result := Copy(str, 1, p - 1);
str := Copy(str, p + Length(cEOL), high(Integer));
end;
end;

function GetAndDelNextWord(var str: String): String;
var e, s : Integer;
begin
s := 1;
while (s <= Length(str)) and (IsWhitespace(str[s])) do INC(s);
e := s;
while (s <= Length(str)) and (not IsWhitespace(str[s])) do INC(s);
result := Copy(str, e, s - e);
while (s <= Length(str)) and (IsWhitespace(str[s])) do INC(s);
str := Copy(str, s, high(Integer));
end;

procedure ReadStuff(x: String; numbers: TDataArray);
var line : String;
numberStr : String;
numberVal : Integer;
begin
while ( Length(x) > 0 ) do
begin
line := GetAndDelNextLine(x);
repeat
numberStr := GetAndDelNextWord(line);
if ( numberStr[1] <> ';' ) then
begin
numberVal := StrToIntDef(numberStr, high(Integer));
if ( numberVal <> high(Integer) ) then
begin
SetLength(numbers, Length(numbers) + 1);
numbers[Length(numbers) - 1] := numberVal;
end;
end
else
line := '';
until ( Length(line) = 0 );
end;
end;

Regards !
Tueblo

Russell Libby

Of course, if raw speed is required, then the following should perform very nicely

Russell

----

type
TIntArray = Array of Integer;

function ParseFileList(FileName: String): TIntArray;
var listData: TList;
lpszData: PChar;
lpszAnchor: PChar;
hFile: THandle;
hMapping: THandle;
dwSize: DWORD;
dwValue: Integer;
dwError: Integer;
szValue: String;
begin

// Create list
listData:=TList.Create;

// Resource protection
try
// Open the file for reading
hFile:=CreateFile(PChar(FileName), GENERIC_READ, FILE_SHARE_READ or FILE_SHARE_WRITE, nil, OPEN_EXISTING, FILE_ATTRIBUTE_NORMAL, 0);
// Check handle
if (hFile <> INVALID_HANDLE_VALUE) then
begin
// Resource protection
try
// Get the file size
dwSize:=GetFileSize(hFile, nil);
// Create the file mapping
hMapping:=CreateFileMapping(hFile, nil, PAGE_READONLY or SEC_COMMIT, 0, 0, nil);
// Check mapping
if (hMapping > 0) then
begin
// Resource protection
try
// Map view of file
lpszData:=MapViewOfFile(hMapping, FILE_MAP_READ, 0, 0, 0);
// Check data
if Assigned(lpszData) then
begin
// Resource protection
try
// While not null
while (lpszData^ > #0) do
begin
// Skip until a char we want
while not(lpszData^ in [#0, ';', '0'..'9']) do Inc(lpszData);
// Check for number
if (lpszData^ in ['0'..'9']) then
begin
// Set anchor
lpszAnchor:=lpszData;
// Parse number
while (lpszData^ in ['0'..'9']) do Inc(lpszData);
// Extract string
SetString(szValue, lpszAnchor, lpszData-lpszAnchor);
// Convert to integer
Val(szValue, dwValue, dwError);
// Check error
if (dwError = 0) then
listData.Add(Pointer(dwValue))
else
listData.Add(Pointer(MaxInt));
end
// Check for comment
else if (lpszData^ = ';') then
begin
// Parse till end of line
while not(lpszData^ in [#10, #13]) do Inc(lpszData);
// Consume end of line
while (lpszData^ in [#10, #13]) do Inc(lpszData);
end;
end;
finally
// Unmap view of file
UnmapViewOfFile(lpszData);
end;
end;
finally
// Close the handle
CloseHandle(hMapping);
end;
end;
finally
// Close the handle
CloseHandle(hFile);
end;
end;
finally
// Resource protection
try
// Convert list to integer array
SetLength(result, listData.Count);
// Move data over
for dwValue:=0 to Pred(listData.Count) do result[dwValue]:=Integer(listData[dwValue]);
finally
// Free list
listData.Free;
end;
end;

end;

Russell Libby

You can also drop the string - > int conversion and perform it manually. Doing so dropped the time on a 3MB file from 140 ms to an average of 38 ms. That time includes creating the resulting integer array.

Russell

---

function ParseFileList(FileName: String): TIntArray;
var listData: TList;
lpszData: PChar;
lpszAnchor: PChar;
hFile: THandle;
hMapping: THandle;
dwSize: DWORD;
dwValue: Integer;
dwMult: Integer;
begin

// Create list
listData:=TList.Create;

// Resource protection
try
// Open the file for reading
hFile:=CreateFile(PChar(FileName), GENERIC_READ, FILE_SHARE_READ or FILE_SHARE_WRITE, nil, OPEN_EXISTING, FILE_ATTRIBUTE_NORMAL, 0);
// Check handle
if (hFile <> INVALID_HANDLE_VALUE) then
begin
// Resource protection
try
// Get the file size
dwSize:=GetFileSize(hFile, nil);
// Create the file mapping
hMapping:=CreateFileMapping(hFile, nil, PAGE_READONLY or SEC_COMMIT, 0, 0, nil);
// Check mapping
if (hMapping > 0) then
begin
// Resource protection
try
// Map view of file
lpszData:=MapViewOfFile(hMapping, FILE_MAP_READ, 0, 0, 0);
// Check data
if Assigned(lpszData) then
begin
// Resource protection
try
// While not null
while (lpszData^ > #0) do
begin
// Skip until a char we want
while not(lpszData^ in [#0, ';', '0'..'9']) do Inc(lpszData);
// Check for number
case lpszData^ of
'0'..'9' :
begin
// Set anchor
lpszAnchor:=lpszData;
// Parse number
while (lpszData^ in ['0'..'9']) do Inc(lpszData);
// Calculate size
dwSize:=lpszData - lpszAnchor;
// Set anchor for backwards walking
Inc(lpszAnchor, Pred(dwSize));
dwValue:=0;
dwMult:=1;
while (dwSize > 0) do
begin
Inc(dwValue, (Ord(lpszAnchor^) - 48) * dwMult);
dwMult:=dwMult * 10;
Dec(lpszAnchor);
Dec(dwSize);
end;
listData.Add(Pointer(dwValue))
end;
';' :
begin
// Parse till end of line
while not(lpszData^ in [#10, #13]) do Inc(lpszData);
// Consume end of line
while (lpszData^ in [#10, #13]) do Inc(lpszData);
end;
end;
end;
finally
// Unmap view of file
UnmapViewOfFile(lpszData);
end;
end;
finally
// Close the handle
CloseHandle(hMapping);
end;
end;
finally
// Close the handle
CloseHandle(hFile);
end;
end;
finally
// Resource protection
try
// Convert list to integer array
SetLength(result, listData.Count);
// Move data over
for dwValue:=0 to Pred(listData.Count) do result[dwValue]:=Integer(listData[dwValue]);
finally
// Free list
listData.Free;
end;
end;

end;

JonathanH13

ASKER

rllibby - thanks for the elegant solution, it works perfectly, and thanks to others for your contributions.

JonathanH13