--ignore-leading-zeros output columns question

[datatraveller1]

Hi @jqnatividad, first of all, thank you very much for the new --ignore-leading-zero (shortcut -z) option.

It works great, but I have a question. e.g. I use:
qsv joinp -i -z --left-semi --cache-schema=-1 company_id,art_no file1.csv company_id,art_no file2.csv -o left-semi.csv

This works correctly except that the leading zeros are stripped from the output column art_no in left-semi.csv. I had hoped that stripping leadings zeros would only happen for the keys, but if the original value of art_no is 01, the value in the output file gets 1.

Reading the documentation, I'm not sure what is the expected behaviour:
"When set, joins are done ignoring leading zeros. Note that this is only applied to the join keys for both numeric and string columns. The output columns will not have leading zeros."
If "will not have leading zeros" isn't a typo, the behaviour is as expected. If this should read "will not have omitted zeros", it would be a problem. Thank you very much!

[datatraveller1]

I have created an example:

file1.csv

id,company_id,art_no
1,COM1,0724
2,cm2,002
3,com3,78
4,CM2,01
5,Cp5,1
6,CPA,000

file2.csv

id,company_id,art_no
1,com1,724
2,CM2,02
3,Com3,078
4,cm2,1
5,CP5,01
6,cpa,0

command:
qsv joinp -i -z --left-semi --cache-schema=-1 company_id,art_no file1.csv company_id,art_no file2.csv -o left-semi.csv

result left-semi.csv:

id,company_id,art_no
1,com1,724
2,cm2,2
3,com3,78
4,cm2,1
5,cp5,1
6,cpa,""

issues:
The output columns get changed to lower case and leading zeros are stripped. 0 or 000 get changed to an empty string.

expected output:
Same as file1.csv.

[jqnatividad]

Unfortunately, It is working as designed @datatraveller1 .

I'll refactor it so that it creates temporary join columns with all the key transformations and leave the original column untouched. The temporary columns are removed after the join.

WDYT?

[datatraveller1]

Hi @jqnatividad, thank you very much, that sounds great!
--ignore-case (without --ignore-leading-zeros) also leaves the original column untouched, so the new behaviour of --ignore-leading-zeros would be consistent 👍