Using CsvFile.Fields

[pusherman]

Hi all,

I'm using CsvFile to read some quite large CSV files and all is okay apart from CsvFile.Fields.

I notice in the documentation that field names are set to lower case, why is this?

In the files I am using all the field names are capitalised, some are all capitals (acronyms for example), so to have them convert to all lower case is causing a bit of a problem.

Is there a way of converting field names to their original format?

Thanks all
Pusherman

[BruceSteers]

What about CsvFile.KeepNames
https://gambaswiki.org/wiki/comp/gb.uti ... /keepnames

https://gambaswiki.org/wiki/comp/gb.util/csvfile

[pusherman]

Thanks, but at the risk of sounding dumb I can't figure out what format KeepNames takes.

If I say myCSVfile.KeepNames=True then myCSVfile.Fields just returns #0

What am I missing, here?

[BruceSteers]

Thanks, but at the risk of sounding dumb I can't figure out what format KeepNames takes.

If I say myCSVfile.KeepNames=True then myCSVfile.Fields just returns #0

What am I missing, here?

No idea, i've never used CsvFile

according to doc it's just a simple boolean switch.

Property KeepNames As Boolean

Since 3.17

Return or set if field names are kept unchanged or not.

If this property is set, the NoDiacritics property is ignored.

I assume you have 3.17+

It's possible there's a bug but nobody is really using the CsvFile.class so it's not been reported yet.

[pusherman]

Thanks for looking, I'm on the latest version.

I have a work-around that's okay for me. I'm reading the first line of the file with a simple Line Input and substituting what I've read with what .Fields gives me.

Thanks, again.

[cogier]

I think there may be a bug here, as I get the same #0 that you did. But just for fun, here is my workaround.

CSVTest-0.0.1.tar.gz

(8.54 KiB) Downloaded 366 times

[BruceSteers]

It's definitely a bug.

I just submitted a fix.
https://gitlab.com/gambas/gambas/-/merge_requests/338

the code was this...

  For iInd = 0 To $aField.Max
    
    If Not $bKeepNames Then
    
      sField = Trim($aField[iInd])
      
      sField = Replace(sField, String.Chr(160), " ")
      sField = Replace(sField, "\n", " ")
      sField = Replace(sField, "\t", " ")
      
      While InStr(sField, "  ")
        sField = Replace(sField, "  ", " ")
      Wend
      
      sField = String.LCase(sField)
      
    Endif
    
    If $bNoDiacritics Then sField = String.RemoveDiacritics(sField)
    
    If Not sField Then sField = "#" & CStr(iInd)
    $aField[iInd] = sField
  Next

but should be this...

  For iInd = 0 To $aField.Max
    
    sField = Trim($aField[iInd])  ' moved here
      
    If Not $bKeepNames Then
    
      sField = Replace(sField, String.Chr(160), " ")
      sField = Replace(sField, "\n", " ")
      sField = Replace(sField, "\t", " ")
      
      While InStr(sField, "  ")
        sField = Replace(sField, "  ", " ")
      Wend
      
      sField = String.LCase(sField)
      
    Endif
    
    If $bNoDiacritics Then sField = String.RemoveDiacritics(sField)
    
    If Not sField Then sField = "#" & CStr(iInd)
    $aField[iInd] = sField
  Next

The sField was never set if using KeepNames so i moved it to before the If condition.

I've attached the fixed CsvFile.class if you fancy testing it works

[BruceSteers]

Never mind testing the fix Ben has already accepted it so just download latest commit once the gitlab-ci has compiled it all.

[BruceSteers]

That fix went straight into stable 3.19.1 so it's fixed now for latest master and stable

But...
If you may need your program to work on older gambas versions than 3.19.1 then do not use CsvFile.class from gb.util but place the CsvFileFix.class i uploaded into your programs source folder (rename it if you want) and use that class instead.

[pusherman]

Yes, the update works fine for me.

Thanks for your help, guys.