This may be a bug/feature of the json library this tool uses, but my CSV had a large number in one of the columns, and it was evidently parsed as different number (also negative). I'm guessing my number was larger than whatever type json2csv's json parse was expecting:

$ echo '{"x": 110750513297351875876238378613844499030}' |  /opt/gocode/bin/json2csv -k x
-9223372036854775808

Hi,

I don't have go, and have no plans of installing/using go in my work place.

Can you please add a pre-compiled json2csv version without go, and/or any run time dependency. I'd like to use this tool.

Thank you.

The following doesn't work

echo '{"FOO": {"BAR.BAR": "BURGER"}}' | json2csv -k FOO.'BAR.BAR' #or FOO.BAR.BAR

It'd be nice to be able to do something like json2csv -k * where the * indicates all fields from the json object will be included in the output.

json2csv uses a lot of CPU, but only one core. json.Unmarshal seems to be taking up most of the time so I made a forked version that launches a whole bunch of them in parallel

A 207Mb json file with 241804 lines takes 15s to convert to csv with the single-threaded version, using all cores (2 real + 2 hyperthreading) it takes 7s. However using just one core on the multi-threaded version takes 17s It might be worth to keep the single-threaded version in the code and use that when running on one core.

parallel version

https://superuser.com/questions/1412477/how-can-i-view-my-instagram-backup-data-in-a-friendly-way

I have json file that looks like below. json2csv works if I make this each entry a single line. Is this format it fails. I there a way for json2csv handle file formatted in this way?
{
“count”:10,
“entries”:
[
{
“id”:20232913,
“application”:“AuditExampleLogin1”,
“user”:“TestUser”,
“time”:“2018-08-15T23:59:40.186-07:00”,
“values”:
null
},
{
“id”:20232914,
“application”:“AuditExampleLogin1”,
“user”:“AUser”,
“time”:“2018-08-15T23:59:55.186-07:00”,
“values”:
null
},

attempting to convert a json using the instructions on the readme, I get error: unknown option-k'`

This occurs even when using the exact data and command written on the readme, i.e., input.json:

[
	{"user": {"name":"jehiah", "password": "root"}, "remote_ip": "127.0.0.1", "dt" : "[20/Aug/2010:01:12:44 -0400]"},
	{"user": {"name":"jeroenjanssens", "password": "123"}, "remote_ip": "192.168.0.1", "dt" : "[20/Aug/2010:01:12:44 -0400]"},
	{"user": {"name":"unknown", "password": ""}, "remote_ip": "76.216.210.0", "dt" : "[20/Aug/2010:01:12:45 -0400]"}
]

json2csv -k user.name,remote_ip -i input.json -o output.csv

error: unknown option `-k'

Trying to test this and I am unable to get anything to output into my csv file through powershell.

I created a file with the JSON data from the README just copy/pasted (test2.json) and when running

json2csv -k user.name,remote_ip -i test2.json -o output.csv

it hangs and then gives the error "ERROR Decoding JSON at line 1: unexpected end of JSON input." I've attached the data as a text file for reference.

test2.txt

json2csv currently turns null values to "", which is not always desirable

I'm trying to parse through Instagram follower data, structured as such:

{
    close_friends: { 
        username: timestamp
    }
    followers: {
        follower1: timestamp1, 
        followerN: timestampN
    }
}

Using the command ./json2csv -i input.json -k followers -o output.csv outputs all followers+timestamp in a single row/cell. Is there a way to output 1 follower per row?

Thank you.

When running make:

Error received:

Updating goal targets....
File all' does not exist. Filejson2csv' does not exist.
Must remake target `json2csv'.
cc -I. -I/usr/local/include -O2 -g -o json2csv json2csv.c -L. -L/usr/local/lib -ljson
json2csv.c:156:15: error: non-void function 'parse_fields' should return a value [-Wreturn-type]
if (!str) return;
^
1 error generated.
make: *** [json2csv] Error 1

I have done the json-c pre-requisite changes.

please email me at [email protected] for more info

Thanks!

Thanks for your fantastically useful json2csv script - I've been using it to parse data from OpenLibrary dumps. It's working very well, even though the OL data is very inconsistently structured. One question, though, if I may...

In a case where there are commas within an item, eg

{"subjects": ["Books and reading -- Fiction.", "Storytelling -- Fiction.", "Death -- Fiction.", "Jews -- Germany -- History -- 1933-1945 -- Fiction."]}

json2csv appears to strip out the commas within the value, so the four different subjects all get merged into one. It comes out like this for -k subjects:

[Books and reading -- Fiction. Storytelling -- Fiction. Death -- Fiction. Jews -- Germany -- History -- 1933-1945 -- Fiction.]

Is there a straightforward way to get it to preserve those multiple items within a value? (I don't need them as separate fields in the CSV, but would like to preserve the distinction within the 'subjects' field, if you see what I mean - so they could be delimited by something other than a comma.)

(I tried using the -d flag to set a different field delimiter, e.g. semicolon, but it still stripped out the commas as above.)

Edit: another example...
"subject_places": ["United States", "China"]
comes out as
[United States China]
so it's not really practical to find some automated way of parsing that alas.

If I have the source JSON

{"message": "this is a \"test quote\""}

then when I run json2csv -k message <JSON, the output looks like

"this is a "test quote""

which breaks any CSV reader. The escaping used in the JSON should be carried over to the CSV.

Trying to convert JSON which is an array rather than a single object and I get back the following error:

ERROR Decoding JSON at line 1: json: cannot unmarshal array into Go value of type map[string]interface {}

Source of data https://pressbooks.online.ucf.edu/wp-json/pressbooks/v2/books/

Hi!!

I've installed json2csv by

$go get github.com/jehiah/json2csv

but when I try to do

$cat someFile.json | json2csv -k time,tx_index > output.csv

then I get the following:

2015/04/28 17:18:20 ERROR Decoding JSON at line 1: unexpected end of JSON input
{
2015/04/28 17:18:20 ERROR Decoding JSON at line 2: invalid character ':' after top-level value
"time": 1348310820,
2015/04/28 17:18:20 ERROR Decoding JSON at line 3: invalid character ':' after top-level value
"tx_index": 9741630
2015/04/28 17:18:20 ERROR Decoding JSON at line 4: invalid character '}' looking for beginning of value
}

and so on... What am I missing?

Thank you for you time.

When data is unable to be parsed, the bytes are output which isn't helpful when you want some human readable output of what the record was.

[anna@ ~]$ zcat /stream_archive/decodes.2012-06-10_11.log.gz | json2csv -k g | uniq > ~/test.cav
ERR: unable to parse json ({"a": "Mozilla/5.0 (Windows NT 5.1; rv:2.0.1) Gecko/20100101 Firefox/4.0.1", "c": "ID", "nk": 0, "tz": "Asia/Jakarta", "gr": "08", "g": "J5BebP", "i": "...", "h": "J5BebO", "k": "4fd4855a-00384-06073-271cf10a", "l": "mekaputra", "hh": "bit.ly", "r": "Dy\u00d9\u000b\uff8c\u00e1\uffb1\u0005<\uff95\u001e\fP\u00cd\u00df\u000b\u00d0\u00cd\u00dc\u0002\uff90\u0002\u0004\f\u012eo\u000b\u0010\u00c9\u0015\u0011P\u00c9\u0015\u0011", "u": "http://livebeta.kaskus.us/post/000000000000000694762851#post000000000000000694762851", "t": 1339327834, "hc": 1337025892, "kw": "movenkaskusbdg", "cy": "Malang", "ll": [-7.9797000885009766, 112.63040161132812]}
)

Does your code handle nested indices in json files?
I have a file that has "user" as an index, which has a nested index called "lang". "user.lang" doesn't seem to work, so I was wondering if there's any way to save "lang" in the csv.

I'm sorry for asking a basic question but I'm new to golang. Running go get github.com/jehiah/json2csv on my CentOS 7 finished without any error but json2csv resulted in the error, bash: json2csv: command not found.... Please suggest where am I going wrong.
P.S. I had to reset GOROOT to allow installation to end successfully as suggested on this page.

(or to make it conditional)

You know the drill

I love this tool! I'm using it to convert json files from Twitter's streaming API to csv. I have very little programming skills, so this tool is really good for me.

However I'm stuck at this point where I would like to extract values from the coordinate pairs within tweets. I think the coordinates are stored as an array:

"coordinates": {"coordinates": [3.82955502, 51.48484907], "type": "Point"}

Would it be possible to store the X and Y-coordinate in a separate csv field?

I tried: "json2csv -k coordinates.coordinates[0],coordinates.coordinates[1] -i ...etc" but this doesn't work.

Thanks in advance!
Roeland

[
[1508834700000,"5688.5","5735.4","5676.51","5699.4","9.9216"],[1508835600000,"5699.4","5727","5699.4","5727","5.0476"],[1508836500000,"5727","5733.6","5644.31","5693.94","66.833"],
]

like this....
how to convert?
thx...