--fixlengths option

[seamusabshere]

$ cat a.csv
a,b,c
1,2,3
1

$ scrubcsv a.csv
3 rows (1 bad) in 0.00 seconds, 106.58 KiB/sec
a,b,c
1,2,3
Too many rows (1 of 3) were bad

it would be nice if i could do

$ scrubcsv --fixlengths a.csv
3 rows (0 bad) in 0.00 seconds, 106.58 KiB/sec
a,b,c
1,2,3
1,,

[seamusabshere]

obviously this is duplicative of xsv fixlengths, but it's inconvenient to install and call both binaries for what is properly a CSV scrubbing task

[emk]

I think this is out of scope for scrubcsv, unfortunately. scrubcsv is designed around the idea that incorrect line lengths are generally a sign of buggy CSV writers that can't get quoting right, and scrubcsv's most fundamental job is to throw out those lines.

If you know that your CSV is 100% valid but has variable-length lines, then try xsv fixlengths first. Since this feature would conflict with the most basic feature of scrubcsv, I feel like it belongs in another tool.

[emk]

I suppose we could consider a mode where short lines were OK, and long lines weren't, but there's a good chance you'd need to know exactly how each specific input file was broken.