[feature] Line masking (ignoring lines or part of lines in matching but displayed in output) about ripgrep HOT 3 CLOSED

Shell pipelines are your best bet. e.g., rg -U 'red stage,\p{any}*foo' | rg 'red|foo'.

Otherwise, your task seems pretty specialized. If you need to do one or two ad hoc queries, then ripgrep is probably a good fit. But if you need something more systematic and flexible, I'd probably build a bespoke tool for it personally.

from ripgrep.

Actually I misunderstood what multiline does, I want to match phrases across line breaks something like "red\s\stage\sfoo" but that's easy to do and okay (if you put aside the interleaved metadata problem). I think your example doesn't work because it would match 'red stage, blue stage, foo'.

I think what I'll end up doing is pre-processing the file, removing the metadata, extracting the matching line numbers from rg and printing those lines in the original file using sed/awk.

(I understand that it's probably too far afield of what rg is intended to do, but what I was originally thinking was that you would be able to, in the most general case, do something like rg --extractor-fn "jq '.text'" to be able to search only the text across lines (and still output the entire data structure) of something like {'ts': 111, 'text': 'this text continues on'} {'ts':112, 'text': ' the next line'}. Or for simpler cases that can be handled with regex: rg --mask "-- \d\d:\d\d$")

Anyway thanks and feel free to close.

from ripgrep.

Yeah ripgrep isn't arbitrarily flexible like that. With that said, it does have the --pre flag, which lets you run any program to transform the input in whatever way you want before ripgrep will search it. The docs give an example of how to use ripgrep to transparently search the plain text of PDF files.

from ripgrep.