pd.DataFrame.to_dict('records') increase precision unexpectedly

[tashrifbillah]

When a Pandas DataFrame is converted into a Python dictionary:

dpimport/dppylib/__init__.py

Line 141 in 3127f6f

data_blob.extend(chunk.to_dict('records'))

precision of some numbers increase unexpectedly (0.999-->0.999000000001):

There is also a post on stackoverflow about it.

---

Tashrif's solution:

chunk.round(5).to_dict('records')

[tashrifbillah]

@hre80 supports this solution

[tashrifbillah]

Changes are in /data/predict/dpimport, will be pushed up once found stable.

[tashrifbillah]