Calling fields by ID

[masgo]

Is there a way to call fields by id or by name within the same call? For example, I would like to do something like then rename 1,newName. This should rename column 1 in "newName".

I saw positional field names, but it seems they only work in the DSL, i.e. when called like $[[[3]]]

I am usually dealing with tsvlite or csv files. Sometimes they have long and complicated column names, sometimes even with non-english letters and other special characters etc. (example: ID/Eintrag gelöscht (001)) Miller works fine with that, but it can be quite difficult to address these fields, from within a script. I know that I can use curly braces or " (e.g., "ID/Eintrag gelöscht (001)" ). But now the charset of the bash script file and/or of the console becomes important. This is painful.

In these cases I use --implicit-csv-header --headerless-csv-output and adress the columns by id. Works fine, but longer scripts become hard to read. Using something like then rename [[[1]]],newName at the beginning would make life easier. But since this is outside of the DSL, miller searches for a field called [[[1]]]. Is there a way to make it search for the first column?

[aborruso]

I also think it can be useful #310

[masgo]

Thanks, I have not seen the ticket. Your answer also answers my question. It's not possible right now.

[johnkerl]

@masgo @aborruso this is, at essence, the same issue as in #763.

Namely, the field names for verbs are just strings with no syntactic structure. In #763 if you have a field item which is a map with subfield name then in the DSL you can refer to item.name which will index $* with "item", then index that with "name", where the "." is parsed as a delimiter. But the cut verb (and all non-DSL verbs) looks for a single field named "item.name" (the "." is not parsed out as a delimiter).

Here, similarly, the DSL parses the [[ and ]] syntactically to give you the field indexed by 3. But non-DSL verbs look for (and fail to find) a field whose name is literally "[[3]]".

What we need is a grammar, and a parser, for field "names" in non-DSL verbs to give them the same kind of expressiveness they already enjoy in the DSL.