CSV splitter

csv-splitter.sh

#!/bin/bash
#split csv file with header into separate files based on key column (input should be sorted by key column)

if [ $# -eq 0 ]; then
    echo "Usage: $0 <input_csv_file>"
    exit 1
fi

INPUT_CSV="$1"

current_key=""
output_file=""

read -r header <<< "$(head -n 1 "$INPUT_CSV")"
# IFS sets splitter as comma, then read takes column 1, column 2 as 'key' and 'rest' as unseparated values
# so for 1st column it would be "read key rest"
# and for 2nd "read col1 col2 key rest"
# with -r, the backslashes are treated as literal characters, preserving them in the output
while IFS=, read -r col1 key rest; do
    if [[ $key == "" ]]; then
        continue
    fi
    
    if [[ $key != $current_key ]]; then
        
        current_key="$key"
        output_file="${current_key}.csv"
        touch "./$output_file"
        echo "$header" >> "./$output_file"
    fi
    
    echo "$col1,$key,$rest," >> "./$output_file"
done < "$INPUT_CSV"

echo "Files have been created based on the first column values."