How to check in the DSL whether a column exists

[halloleo]

I have CSVs like this

mountain,height,
Everest,8849,
Home Hill,100,My name

Now I want to add a better header name for the last column, if it hasn't a good header name yet. I do this with

mlr --csv -N put 'if(NR==1) { if($[3]=="") { $[3]="comment" } }' file.csv

This works - but only when the 3rd column exists. An input file like

mountain,height
Everest,8849
K2,8611 

gives me the error "mlr: conditional expression did not evaluate to boolean."

So how can I check that the 3rd column exists before I check for its header value?

[aborruso]

It could be useful the template verb. Running

mlr --csv template -f mountain,height,comment input.csv

you will have a good result for both your examples.

[halloleo]

Thanks @aborruso! Will try the templateverb.

However is there a way to check for column existence in the DSL?

[aborruso]

However is there a way to check for column existence in the DSL?

Try this

mlr -N --csv put 'if(NR==1) { if(!is_present($3)) { $[3]="comment" } }' then unsparsify input.csv

[halloleo]

Aah, is_present! So cool! Will try both ways later today and report back.

[halloleo]

OK, I tried the template verb, and it is a great tool to know (to unify the headers in mutiple CSVs), but in my case it doesn't work, beacuse I want to keep the number of columns in each CSV as is.

Then I tried the DSL solution with is_present - and it works great! (I didn't use the unparisfy though.) Thanks a lot for your help, @aborruso!

PS: I don't know whether I should mark this answer thread as solution, because the solution evolved for me from the comments not from the initial answer which marking would highlight. - Maybe we leave it as is...

[aborruso]

A note. In this

mountain,height
Everest,8849
K2,8611

the third field does not exist.

Instead here

mountain,height,
Everest,8849,
Home Hill,100,My name

it exists. it has no label, but it exists.

[halloleo]

Yes, indeed that's how the input files can differ.

And the output should reflect this too: The number of existing columns should stay as in the input file.

[aborruso]

@halloleo I'm cloning a reply to map a "right" answer.

To check for column existence in the DSL, you could use is_present function:

mlr -N --csv put 'if(NR==1) { if(!is_present($3)) { $[3]="comment" } }' input.csv