PERF: improved perf in .to_json when lines=True

[jreback]

closes #14408

master

In [1]: np.random.seed(1234)

In [2]: N = 100000
   ...: C = 5
   ...: df = DataFrame(dict([('float{0}'.format(i), np.random.randn(N)) for i in range(C)]))

In [3]: %timeit df.to_json('foo.json',orient='records',lines=True)
1 loop, best of 3: 3.67 s per loop

PR (with proper encoding/decoding, IOW work directly on the bytes and minimize copies)

In [5]: %timeit df.to_json('foo.json',orient='records',lines=True)
10 loops, best of 3: 137 ms per loop

[jreback]

xref #14391

cc @joshowen

[codecov-io]

Current coverage is 85.26% (diff: 100%)

Merging #14429 into master will decrease coverage by <.01%

@@             master     #14429   diff @@
==========================================
  Files           140        140          
  Lines         50639      50622    -17   
  Methods           0          0          
  Messages          0          0          
  Branches          0          0          
==========================================
- Hits          43177      43161    -16   
+ Misses         7462       7461     -1   
  Partials          0          0          

Powered by Codecov. Last update 286b9b9...e855e1f

[jorisvandenbossche]

Looks good to me

[jreback]

closed by 7cad3f1