jq (or alternatives) convert separated file into rows of json also prepending fields

Question

I have a file (file.txt) like:

field1|field2|field3 
field1|field2|field3 
field1|field2|field3 

(actually something like 83 fields, 200k rows) I need to convert it into something like:

field3.field2|{"myname1":"field1","myname2":"field2","myname3":"field3"}
field3.field2|{"myname1":"field1","myname2":"field2","myname3":"field3"}
field3.field2|{"myname1":"field1","myname2":"field2","myname3":"field3"}

I'm currently doing it in several steps, but it takes ages:

cat file.txt |jq -nR '[inputs | split("|") | {
"myname1": .[0],
"myname2": .[1],
"myname3": .[2]}]' > /var/tmp/tempfile.txt

cat /var/tmp/tempfile.txt | jq -c '.[]' |while read row;do
key=$(echo "$row"|jq -r '(.myname3)+"."+(.myname2)')
echo "$key|$row"
done

is there a better way with jq or something else (under Linux) to achieve the same goal ?

thanks

Answer 1

Here's a solution using jq with string interpolation and conversion to JSON using tojson:

jq -Rr './"|" | "\(.[2]).\(.[1])|\({
  "myname1": .[0],
  "myname2": .[1],
  "myname3": .[2]
} | tojson)"' file.txt

field3.field2|{"myname1":"field1","myname2":"field2","myname3":"field3"}
field3.field2|{"myname1":"field1","myname2":"field2","myname3":"field3"}
field3.field2|{"myname1":"field1","myname2":"field2","myname3":"field3"}

Demo