Repository navigation
Extend json.tool to handle jsonlines (with a flag) #75734
Description
| BPO | 31553 |
|---|---|
| Nosy | @rhettinger, @vstinner, @ezio-melotti, @serhiy-storchaka, @lisroach |
| PRs |
Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.
Show more details
GitHub fields:
assignee = 'https://github.com/lisroach'
closed_at = <Date 2018-11-07.10:25:30.301>
created_at = <Date 2017-09-22.16:41:06.565>
labels = ['3.8', 'type-feature', 'library']
title = 'Extend json.tool to handle jsonlines (with a flag)'
updated_at = <Date 2018-11-07.10:25:30.300>
user = 'https://bugs.python.org/EricMoyer'bugs.python.org fields:
activity = <Date 2018-11-07.10:25:30.300>
actor = 'serhiy.storchaka'
assignee = 'lisroach'
closed = True
closed_date = <Date 2018-11-07.10:25:30.301>
closer = 'serhiy.storchaka'
components = ['Library (Lib)']
creation = <Date 2017-09-22.16:41:06.565>
creator = 'Eric Moyer'
dependencies = []
files = []
hgrepos = []
issue_num = 31553
keywords = ['patch']
message_count = 9.0
messages = ['302755', '302781', '302783', '302785', '302904', '302953', '328290', '328396', '329411']
nosy_count = 6.0
nosy_names = ['rhettinger', 'vstinner', 'ezio.melotti', 'serhiy.storchaka', 'lisroach', 'Eric Moyer']
pr_nums = ['3846', '10051']
priority = 'normal'
resolution = 'fixed'
stage = 'resolved'
status = 'closed'
superseder = None
type = 'enhancement'
url = 'https://bugs.python.org/issue31553'
versions = ['Python 3.8']
Activity
EricMoyer commentedon Sep 22, 2017
json.tool should have the ability to format jsonlines data. It is a very commonly used format in many industries. Right now when given jsonlines (for example):
Works. But:
Reports an error: "Extra data: line 2 column 1 - line 3 column 1 (char 58 - 100)"
I propose that the above behavior be retained but:
Should print:
{
"ingredients": [
"frog",
"water",
"chocolate",
"glucose"
]
}
{
"ingredients": [
"chocolate",
"steel bolts"
]
}
If someone else agrees this is an appropriate enhancement, I can start work on a PR in a couple of weeks.
rhettinger commentedon Sep 23, 2017
This seems like a reasonable enhancement.
serhiy-storchaka commentedon Sep 23, 2017
Jsonlines is an external to JSON format. You should split the data on lines and pass every line to the JSON parser separately. The same you should do with json.tool.
ezio-melotti commentedon Sep 23, 2017
I agree with Raymond that this is a reasonable request.
Even if jsonlines is not part of the JSON specification, the format is quite common, and practicality beats purity (espcially considering that we are just talking about json.tool).
rhettinger commentedon Sep 25, 2017
Eric, did you want to write the PR yourself or would you like for Lisa to bring it to fruition? If you're going to do it yourself, Lisa will serve as the primary reviewer (with either Ezio or me doing the final sign-off).
EricMoyer commentedon Sep 25, 2017
I planned to write the PR myself, on the principle of "it's my itch, I
should scratch it." But if you think it is better for someone else to write
it, you know the codebase better than I, and you know how long a PR takes
to review/fix. My intention is to help. I'll do whatever is easier for you.
On Sep 24, 2017 10:01 PM, "Raymond Hettinger" <[email protected]>
wrote:
serhiy-storchaka commentedon Oct 23, 2018
The output is not a valid JSON format, and is not a valid JSON Lines format. What you are going to do with it?
vstinner commentedon Oct 24, 2018
It looks useful to me to read a nicely indented JSON in the terminal (without specifying an output file). It's more readable that compact JSON on a single line.
But should we add an option to write the output on a single line? It would be the opposite of the tool description:
"A simple command line interface for json module to validate and *pretty-print* JSON objects."
serhiy-storchaka commentedon Nov 7, 2018
New changeset f194479 by Serhiy Storchaka (HongWeipeng) in branch 'master':
bpo-31553: add --json-lines option to json.tool (bpo-10051)
f194479