[SOLVED] Extract unique code and uid from log file using bash script

Issue

I'm new to bash scripting and want to extract code and uid (unique) from log file into new csv file for further processing.

The file contains:

2021-02-16T10:13:54.629Z,app,"Code: 01-87-04 & Diff [{:uid=>""asdf1234"", :sid=>636614}] / [{:uid=>""asdf1234"", :sid=>0}]"
2021-02-16T10:11:22.914Z,app,"Code: 01-33-05 & Diff: [{:uid=>""kx4oa3hv"", :sid=>644448}, {:uid=>""v7jfxsum"", :sid=>643063}] / [{:uid=>""kx4oa3hv"", :sid=>0}, {:uid=>""v7jfxsum"", :sid=>0}]"

The desired output:

01-87-04, asdf1234
01-33-05, kx4oa3hv
01-33-05, v7jfxsum

I started with this script and blocked on extracting the multiple uids. Any help is appreciated.

#!/bin/bash

echo "----->Loading file"
filename="some.csv"
while read -r line; do
  scene_id=`echo "$line" | awk '{print $2}'`

  echo $scene_id
done < "$filename"

Solution

Using any awk and sort in any shell on every Unix box:

$ awk -v OFS=', ' '{n=split($0,uids,/""/); for (i=2;i<n;i+=2) print $2, uids[i]}' file | sort -u
01-33-05, kx4oa3hv
01-33-05, v7jfxsum
01-87-04, asdf1234

Answered By - Ed Morton