Compare files with different columns count

[tmtben]

Hello,

Thanks for this great tool!

Here is my use case:
• my base-csv contains whole rows with many columns.
• my delta-csv contains a list of primary keys (one column only).
I would like to get diff comparing only primary keys.

I got this error with csvdiff version 1.3.0:

# csvdiff base.csv pk.csv
csvdiff: command failed - base-file and delta-file columns count do not match

Best regards,
Ben

[aswinkarthik]

While validating both the configuration files, there is a check for the column cout here

How about

csvdiff base.csv pk.csv --columns 0

This means only one column is compared and it will also check if that index of that column is present in both the CSV files.

[tmtben]

Using version 1.4 with "columns" flag:

# csvdiff base.csv pk.csv --columns 0
csvdiff: command failed - base-file and delta-file columns count do not match

[aswinkarthik]

I was thinking of introducing a feature where if we specify columns flag, we dont need to check if columns count match. All that matters is if the specified column is present on both csvs.

That would satisfy your requirement i believe.

[pascalbe-dev]

Any update on this?

This would be really nice.
Furthermore it would be nice, if we could check which columns have been added (with which values).

[tmtben]

Maybe just replace following line

csvdiff/cmd/config.go

Line 55 in 1007bf3

if baseRecordCount != deltaRecordCount {

by this one
if len(valueColumnPositions) == 0 && baseRecordCount != deltaRecordCount {

Now, I can use "--columns 0" to compare only the first column of two CSV files having different headers.