Output to multiple files

[zxcat]

I need to move the values of some keys of a large JSONs input into separate files. Currently, this requires writing several jq commands for the same JSON, which is very slow with large inputs.

It would be great if jq could read and parse JSON once and then output to different files.
Alternatively, it could set ENV variables so I can write files outside of jq.

[itchyny]

You can bind shell variables by evaluating jq output.

eval "$(jq -nr '{x:1,y:{y:2},z:{"some":"large json"}} | @sh"x=\(.x|tojson); y=\(.y|tojson); z=\(.z|tojson)"')"

echo "$x"
echo "$y"
echo "$z"

[emanuele6]

Someone asked time ago on the #jq IRC channel on libera.chat

I proposed this solution:

#!/bin/sh --

jq -r '
    try (               [ .index, .isActive ] | @tsv "1\(.)"),
    try (.friends[]   | [  .name,      .age ] | @tsv "2\(.)"),
    try (.companies[] | [  .name,  .address ] | @tsv "3\(.)")
' file.jsonl \
| sed -n -e 's/^1//w file1.tsv' -e t \
         -e 's/^2//w file2.tsv' -e t \
         -e 's/^3//w file3.tsv'

N.B. if your output is not @tsv or something else that is guaranteed to be one line long, you should use something like "1\((. / "\n")[])" instead of just "1\(.)".

[zxcat]

Thanx, I've tried both.
The first one is more flexible for me, coz allows to put partial JSON into a variable.