Data Compare error

[stevepjohnson]

I found the new Data Compare for CSV files by accident, and it's really nice!

However, I found a bug in the compare

Although comparing files containing just the line with the problem doesn't show the bug, I still get the error on obfuscated copies of my CSV originals

So attached is a zip of the two obfuscated CSV files. The error is on comparing the 'iiiiiii' line:- this line is identical in both files, but shows as different in cirrus (on windows) - seems in one file the line is correctly split by the commas, but in the other file, for some reason (possibly because of a single apostrophe) the last few commas are not seen. As this doesn't happen for files containing just these lines, it's more an issue of context.

So with this data, the lines showing the problem are displayed 2nd, and are lines 74/75 in each file.

[Erik]

Hi Steve,

The reason you are seeing this behavior is that the apostrophe is automatically being detected as the text qualifier. In the next release, we will change the detection to handle this specific case. For now, you will have to edit your Data Format:

1. Under the "Tools" menu, select "File Formats" to open the dialog.
2. Select "Data Files" from the list.
3. Switch to the "Type" tab.
4. Uncheck "Detect".
5. Change the "Text qualifier" to "None" (or "Quote" which is the standard for CSV).
6. Click "Save".

This will cause any csv files to always use comma as the delimiter and have no qualifier.

[stevepjohnson]

Thanks Erik - Ok, I think that the ' is the crux of the problem. I got that line to show as identical by removing the extra (') chars in each file - there were 3 in one and 2 in the other.

Anyway - I look forward to the new release !