JSON to CSV

json2csv.bash

#!/usr/bin/env bash

# Bash script to convert JSON data using jq.
###############################################################################
# MIT License
# 
# Copyright (c) 2024 Jens Melgard Churchill <jens@churchill.dk>
# 
# Permission is hereby granted, free of charge, to any person obtaining a copy
# of this software and associated documentation files (the "Software"), to deal
# in the Software without restriction, including without limitation the rights
# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
# copies of the Software, and to permit persons to whom the Software is
# furnished to do so, subject to the following conditions:
# 
# The above copyright notice and this permission notice shall be included in all
# copies or substantial portions of the Software.
# 
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
# SOFTWARE.
###############################################################################

# Ensure IBM i QOpensys open source packages are on PATH...
export PATH="/QOpensys/pkgs/bin:$PATH"

err () {
    >&2 printf 'Error %4s: %s\n' "$SYNTAX_ERROR" "$*"
}

info () {
    if [ $SILENT -eq 0 ]
    then
        >&2 printf '%s\n' "$*"
    fi
}

OPTSTRING=":hi:o:rs"

HELP=
dsp_help () {
    if [ -z "$HELP" ]
    then
        info ''
        info 'Parse JSON to CSV, using jq'
        info ''
        info 'Use -h to display this help screen.'
        info 'Use -i "filepath" to specify JSON input file, or direct data to std. in'
        info 'Use -o "filepath" to specify CSV output file, or direct data to std. out'
        info 'Use -r to replace output file, if it exists.'
        info 'Use -s to silence informational output.'
        info ''
        info 'Script supports single object or array of objects'
        info 'Script will produce a header from the first object, and traverse nested objects'
        info ''
        info 'Syntax example:'
        info "$0 -i 'data.json' -o 'data.csv' -r"
        info ''
    fi

    echo "1"
}

json_to_csv () {
    # If input root is a array of objects, including handling nested objects.
    JQ_STMT_1='(. | map(leaf_paths) | unique) as $cols | map (. as $row | ($cols | map(. as $col | $row | getpath($col)))) as $rows | ([($cols | map(. | map(tostring) | join(".")))] + $rows) | map(@csv) | .[]'

    # If input root is a single object, wrap it as an array...
    JQ_STMT_2="[.] | $JQ_STMT_1"
    
    jq -r "if type==\"array\" then $JQ_STMT_1 else $JQ_STMT_2 end" <&0
}

INPUT_FILE=
OUTPUT_FILE=
REPLACE_OUTPUT_FILE=0
SILENT=0

SYNTAX_ERROR=0

if [ $# -eq 0 ]
then
    SYNTAX_ERROR=-1
    HELP=$(dsp_help)
fi

# Setup using input arguments...
while getopts ${OPTSTRING} OPT
do
case ${OPT} in
    h)
    SYNTAX_ERROR=-2
    HELP=$(dsp_help)
    ;;
    i)
    INPUT_FILE="${OPTARG}"
    ;;
    o)
    OUTPUT_FILE="${OPTARG}"
    ;;
    r)
    REPLACE_OUTPUT_FILE=1
    ;;
    s)
    SILENT=1
    ;;
    :)
    SYNTAX_ERROR=255
    HELP=$(dsp_help)
    err "Option -${OPTARG} requires an argument"
    ;;
    ?)
    SYNTAX_ERROR=1
    HELP=$(dsp_help)
    err "Invalid option: -${OPTARG}"
    ;;
esac
done

# Confirm jq is installed and on PATH...
if ! command jq --version &>/dev/null; then
    SYNTAX_ERROR=10
    err "Required open source command jq not found, please install using yum install jq"
fi

# If we have a regular std. in,
# attempt to use positional arguments for input-file, if unset,
# and validate afterwards...
if [ -t 0 ]
then
    for ARG in "$@"
    do
        [ -n "$INPUT_FILE" ] && break
        
        if [[ "$ARG" != -* ]] && [ -f "$ARG" ]
        then
            INPUT_FILE="$ARG"
        fi
    done

    # Std. in is not being redirected,
    # and input-file does not exist...
    if ! [ -f "$INPUT_FILE" ]
    then
        if [ -n "$INPUT_FILE" ]
        then
            SYNTAX_ERROR=100
            err "Input JSON file \"$INPUT_FILE\" not found, please review input argument -i"
        else
            SYNTAX_ERROR=110
            err "Input JSON file not specified, please review input argument -i"
        fi
    fi
fi

# If we have a regular std. out, and input-file is set,
# attempt to use positional arguments for output-file, if unset,
# and validate afterwards...
if [ -t 1 ] && [ -f "$INPUT_FILE" ]
then
    for ARG in "$@"
    do
        [ -n "$OUTPUT_FILE" ] && break
        
        if [[ "$ARG" != -* ]] && [ "$ARG" != "$INPUT_FILE" ]
        then
            OUTPUT_FILE="$ARG"
        fi
    done
fi

# Std. out is not being redirected,
# use specified output-file and replace it if so instructed...
if [ -t 1 ]
then
    if [ -f "$OUTPUT_FILE" ] && [ 0 -eq "$REPLACE_OUTPUT_FILE" ]
    then
        SYNTAX_ERROR=200
        err "Output CSV file \"$OUTPUT_FILE\" exists, please review output argument -o or specify -r to replace"
    elif [ -n "$OUTPUT_FILE" ]
    then
        exec >"$OUTPUT_FILE"
    fi

# Fail with error that output-file is specified, but std. out is being redirected...
elif [ -n "$OUTPUT_FILE" ]
then
    SYNTAX_ERROR=210
    err "Output file declared, but std. out is being redirected."
fi

info "Command: $0 $*"
info "Input:   ${INPUT_FILE:--stdin-}"
info "Output:  ${OUTPUT_FILE:--stdout-}"

# Are there any detected errors?
# And should they trigger a none zero exit code?
# Below zero is warnings, above is errors...
if [ 0 -ne $SYNTAX_ERROR ]
then
    info "Aborted."
    
    if [ 0 -gt $SYNTAX_ERROR ]
    then
        exit 0
    else
        exit $SYNTAX_ERROR
    fi
fi

info "Converting..."

# Run conversion, either from input-file or std. in...
if [ -t 0 ]
then
    json_to_csv <"$INPUT_FILE"
else
    json_to_csv <&0
fi

This script converts arbitrary JSON into CSV

1. It does this, by using the jq command
2. It can receive and output data both using stdin and stdout, or using input and output file arguments.
3. It should be compatible with QShell on IBM i^tm

Type json2csv.bash -h to view help.

Parse JSON to CSV, using jq

Use -h to display this help screen.
Use -i "filepath" to specify JSON input file, or direct data to std. in
Use -o "filepath" to specify CSV output file, or direct data to std. out
Use -r to replace output file, if it exists.
Use -s to silence informational output.

Script supports single object or array of objects
Script will produce a header from the first object, and traverse nested objects

Syntax example:
./json2csv.bash -i 'data.json' -o 'data.csv' -r

Error  110: Input JSON file not specified, please review input argument -i
Command: ./json2csv.bash -h
Input:   -stdin-
Output:  -stdout-
Aborted.

Here are some examples, they all use the -s option, to suppress verbose output,
note that this verbose output is sent to stderr, so piping stdout should only ever produce result data if conversion succeeds...

Convert single object...

command:
./json2csv.bash -si data/object.json

data:

{
    "firstname": "Jens",
    "lastname": "Churchill"
}

output:

"firstname","lastname"
"Jens","Churchill"

Convert array of objects...

command:
./json2csv.bash -si data/array.json

data:

[
{
    "firstname": "Jens",
    "lastname": "Churchill"
},
{
    "firstname": "John",
    "lastname": "Doe"
}
]

output:

"firstname","lastname"
"Jens","Churchill"
"John","Doe"

Convert nested object...

command:
./json2csv.bash -si data/nested_object.json

data:

{
    "firstname": "Jens",
    "lastname": "Churchill",
    "info": {
        "title": "Author",
        "commits": 1
    }
}

output:

"firstname","info.commits","info.title","lastname"
"Jens",1,"Author","Churchill"

Convert array of nested objects...

command:
./json2csv.bash -si data/array_nested_objects.json

data:

[
{
    "firstname": "Jens",
    "lastname": "Churchill",
    "info": {
        "title": "Author",
        "commits": 1
    }
},
{
    "firstname": "John",
    "lastname": "Doe",
    "info": {
        "title": "Contributer",
        "commits": 0
    }
}
]

output:

"firstname","info.commits","info.title","lastname"
"Jens",1,"Author","Churchill"
"John",0,"Contributer","Doe"

Convert array of nested objects, with incomplete data...

command:
./json2csv.bash -si data/array_nested_objects_incomplete.json

data:

[
{
    "firstname": "Jens",
    "lastname": "Churchill",
    "info": {
        "title": "Author",
        "commits": 1
    }
},
{
    "firstname": "John",
    "lastname": "Doe"
}
]

output:

"firstname","info.commits","info.title","lastname"
"Jens",1,"Author","Churchill"
"John",,,"Doe"

IBM i is registered trademark of the IBM Corporation