CODEWITHSUNDEEP
linuxbashshellsed
hruday
hruday

Reputation: 77

Read first line in a file and append it to first occurrence of the search string and Repeat until EOF

I have a file with following lines in a text file in following format

"abc": "xyz",
"qwe": "uva",
"asd": "lkj",

And a json file in following format

{
    "svn-user-id": "passwd"
    "solution-id": 0,
    "cronos-id": "1",
    "solution-state": "active",
  },
{
    "svn-user-id": "passwd"
    "solution-id": 1,
    "cronos-id": "1",
    "solution-state": "active",
  },
{
    "svn-user-id": "passwd"
    "solution-id": 2,
    "cronos-id": "1",
    "solution-state": "active",
  },

Now i want output of json as follows

{
        "svn-user-id": "passwd"
        "solution-id": 0,
         "abc": "xyz",
        "cronos-id": "1",
        "solution-state": "active",
      },
    {
        "svn-user-id": "passwd"
        "solution-id": 1,
         "qwe": "uva",
        "cronos-id": "1",
        "solution-state": "active",
      },
    {
        "svn-user-id": "passwd"
        "solution-id": 2,
         "asd": "lkj",
        "cronos-id": "1",
        "solution-state": "active",
      },

So how can achieve this using bash?

I've tried with following but it only appends last line of the file.

#!/bin/bash
set -x
file=names.txt #file has list of lines as described above
IFS=$'\n'

for l in `cat $file`
        do
        echo $l 
        sed '/"solution-id": 1,/a \'"$l"'' sample.json # json file as described above
        done

Upvotes: 1

Views: 165

Answers (4)

thanasisp
thanasisp

Reputation: 5975

you can use awk to add these lines.

awk 'FNR==NR{a[NR]=$0; next} /solution-id/{$0=$0 "\n\t" a[++i]} 1' lines file

In case the matches in file are more than the lines, above command will print empty lines for the additional matches. In case you need to loop over lines and print a new line for all matches, you have to make i restart from the beginning, for example like this:

awk 'FNR==NR{a[NR]=$0; next} /solution-id/{$0=$0 "\n\t" a[i++%length(a)+1]} 1' lines file

Note: as already commented, your input is not a valid JSON, probably you have extracted it with text processing from a pretty JSON file. If you need to treat your initial input as JSON you have to use a JSON processing tool.

Upvotes: 1

PiedPiper
PiedPiper

Reputation: 5785

This will do it:

#!/bin/bash
#set -x
IFS=$'\n'

declare -a names
names=(`cat names.txt`)

for l in `cat sample.json`
do
    echo $l
    echo $l | grep -P -q solution-id
    if [ $? -eq 0 ]
    then
        echo "    ${names[$i]}"
        let "i+=1"
    fi
done

Upvotes: 1

potong
potong

Reputation: 58473

This might work for you (GNU sed):

sed 's/^/    /' textFile | sed '/solution-id/R /dev/stdin' jsonFile

Prepend 4 spaces to each line of the text file and pipe the results to the second invocation of sed which appends the lines (one at a time) to the json file when matching on the string solution-id.

Upvotes: 0

Nebril
Nebril

Reputation: 3273

If your solution-id identifiers match the order with your file here is the code that does what you need:

#!/usr/bin/env bash

x=0
cp sample.json tmp.json
while read l
    do
    sed -i "/\"solution-id\": ${x},/a ${l}" tmp.json
    x=$(($x + 1))
    done < names.txt

cat tmp.json

It copies your input file and performs sed replacements in place and prints it at the end.

Upvotes: 0

Related Questions