|
To me I think it makes sense to CATCH this error, though you may have to do a fair amount of diagnosis after you CATCH it. The best way to handle it may depend on what follows this enormous "string". Thus, you can have a loop like this:
openr, lun, inputFile, /GET_LUN
nLines = file_lines(inputFile)
fileText = strarr(nLines)
currentLine = ''
currentLineNo = 0
bytePosAfterLastRead = 0
errorCode = 0
for i = 0, nLines-1 do begin
catch, errorCode
if errorCode ne 0 then begin
catch, /CANCEL
if !error_state.code eq 'IDL_M_FILE_READERR' then begin
print, 'READF error occurred at byte position ' + $
strtrim(bytePosAfterLastRead, 2)
break
endif else print, !error_state.msg
endif else begin
readf, lun, temp
point_lun, -lun, bytePosAfterLastRead
fileText[currentLineNo++] = temp
endelse
endfor
free_lun, lunThe above loop will exit when it encounters the first READF error. I do not know how to keep the loop continuing because I do not think that IDL knows where the end of the extremely long text line is. Thus, when the loop exits in error you should have:
a) the byte location of the first byte in the line that triggered the READF error
b) the line number of the line that triggered the READF error
c) all the text up to the line that triggered the READF error in an IDL string array
When the loop exits you can proceed to read the rest of the file with READU using this approach:
fileInfo = file_info(inputFile)
nRemainingUnreadBytes = fileInfo.size - bytePosAfterLastRead
remainingUnreadBytes = bytarr(nRemainingUnreadBytes)
openr, lun, inputFile, /GET_LUN
point_lun, lun, bytePosAfterLastRead
readu, lun, remainingUnreadBytesAt this point the question remains: How do you parse the 'remainingUnreadBytes' byte array. If there are INDEED orderly text lines after the extremely long string, then you could search the byte array for the first newline character - STRING(10B) in IDL. That would help you find the byte location of the next orderly text line. You could then access the next orderly text lines either in your 'lun' file stream with a new POINT_LUN call, or in your byte array where you could parse out all your text strings by searching for newline chars and converting byte subarrays into strings with IDL's STRING function.
James Jones
|