BUG: Fix to_json lines with escaped characters #15117
Conversation
Current coverage is 84.75% (diff: 100%)
|
pandas/lib.pyx
Outdated
| """ | ||
| cdef: | ||
| Py_ssize_t i = 0, num_open_brackets_seen = 0, in_quotes = 0, length | ||
| Py_ssize_t is_escaping = 0 |
There was a problem hiding this comment.
make in_quotes and is_escaping of type bint
|
pls add a whatsnew note (0.20.0 bug fix section) |
rouzazari
force-pushed
the
to_json_lines_with_escaping
branch
from
60bb9da to
92f4e01
Compare
|
Thanks, @jreback. I made the changes you requested, rebased to master, and squashed the commits. Please note: I revised the test to include a test for escaped characters in keys and values (i.e. columns and data of dataframe). Previously I was only testing escaped characters in values/data. |
doc/source/whatsnew/v0.20.0.txt
Outdated
| - Require at least 0.23 version of cython to avoid problems with character encodings (:issue:`14699`) | ||
| - Bug in converting object elements of array-like objects to unsigned 64-bit integers (:issue:`4471`, :issue:`14982`) | ||
| - Bug in ``pd.pivot_table()`` where no error was raised when values argument was not in the columns (:issue:`14938`) | ||
| - Bug in ``.to_json()`` where lines=True and contents (keys or values) contain escaped characters (:issue:`15096`) |
There was a problem hiding this comment.
use double-backticks around lines=True
pandas/io/tests/json/test_pandas.py
Outdated
| self.assertEqual(result, expected) | ||
| assert_frame_equal(pd.read_json(result, lines=True), df) | ||
|
|
||
| # GH15096 |
There was a problem hiding this comment.
add a 1-line comment on what you are testing here
|
lgtm, minor comments and some linting issues to fix. ping on green. |
Updates existing to_json methodology by adding is_escaping variable, which ensures escaped chars are handled correctly. - Includes test for escaped characters in keys and values (i.e. columns and data). - Includes bug fix in whatsnew - Revised type of in_quotes and is_escaping to bint xref pandas-dev#14693 xref pandas-dev#15096
rouzazari
force-pushed
the
to_json_lines_with_escaping
branch
from
92f4e01 to
d114455
Compare
|
thanks! nice concise PR! keep em coming |
|
Thanks, @jreback. I appreciate the reviews and feedback. FYI: There is no change in runtime performance from the bug fix. This is somewhat obvious, but it was a good exercise for me, so I thought I'd share. I reran the following performance tests from #14408: Master: This branch: |
|
@rouzazari great! note that we already have some asv's for perf testing, I am not sure how much they cover for json though, you can have a look in pandas/asv/benchmarks/packers.py if you think some cases need more, then please PR! |
|
Got it. I'm not too familiar with this area, but I'll take a look. |
Updates existing to_json methodology by adding is_escaping variable, which ensures escaped chars are handled correctly. xref pandas-dev#14693 closes pandas-dev#15096 Author: Rouz Azari <rouz.azari@gmail.com> Closes pandas-dev#15117 from rouzazari/to_json_lines_with_escaping and squashes the following commits: d114455 [Rouz Azari] BUG: Fix to_json lines with escaped characters
3 participants
Updates existing to_json methodology by adding is_escaping variable,
which ensures escaped chars are handled correctly.
Bug description: A simple check of whether the prior char is a backslash
is insufficient because the backslash may itself be escaped.
A test is also included (previously included in #14693).
xref #14693
xref #15096
git diff upstream/master | flake8 --diff(usedpep8radius master --diffinstead)