CSV2JSON is a Python script which converts CSV files to their JSON counterparts. Of course, this thing can be done within a couple of lines of Python, but I wanted the output to be in a specific format. Traditionally, if you will convert a list object to a JSON object with indentation, you will see that the resulting string has individual elements occupying whole lines.

[
    "a",
    "b",
    "c"
]

I did not want this. Instead I wanted the list representing a row to be nicely stacked in one single line.

["a", "b", "c"]

I also wanted to implement an automatic delimiter detector for CSV files. The traditional Sniffer() class provided with the csv module was unable to guess the delimiter had I used something like ~ or # as a delimiter. So, I wrote my own delimiter guessing algorithm.

Essentially csv2json.py is a command line utility. Here is how you can use it: NOTE: For the last two modes of operations presented below, CSV files must end with .csv extension else this program won't catch them during directory traversal.

1. If you wish to convert a single CSV file to a JSON file:

    python csv2json.py path/to/file1/file1.csv -o path/to/file2/file2.json

2. If you wish to convert all the CSV files present under a directory to their corresponding JSON files in some other directory:

    python csv2json.py -dn path/to/csv/files/ -o path/to/json/files/

Here -d (or --depth) flag denotes the depth till which the first directory should be searched for CSV files.

3. If you wish to convert all the CSV files present under a directory to their corresponding JSON representations, concatenate these, and write them to a single JSON file, then use -c or --concatenate flag:

    python csv2json.py -dn -c path/to/csv/files/ -o path/to/json/file/file.json

You can play with the data given in samples directory. I am including a Python script play.py which executes a bunch of commands to give you an idea about how to use this utility:

    python play.py

To clean up after running play.py, you can run the same script with clean flag:

    python play.py clean

If you want to know the thought-process behind the code, kindly refer to csv2json.web present in sandbox/. This project was basically developed in sandbox directory as a literate program. When the code was ready, it was moved here.