dask / dask / 5769 / 1
53%
master: 53%

DEFAULT BRANCH: master
Ran
Files 90
Run time 4s
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 92.497% (-0.3%) from 92.82%
PYTHON=2.7 NUMPY=1.10.4 PANDAS=0.19.0 COVERAGE='true' PARALLEL='false' TEST_IMPORTS='false' XTRATESTARGS=

push

travis-ci

mrocklin 
Add sql loader (#1181)

* Add sql loader

* Start attempt to use ntile instead of offsets

This is untestable as ntile and row_number are not implemented
in sqlite (but in the target sql servers, e.g., PG, they are).

* Tests pass on postgres

* Explicitly set index

* Remove dross

* Try simple SQL->dataframe loader

* Flake fixes

* Now with SQLalchemy

The code looks a little less tidy, and I don't like the isinstance(, str), but
seems to work well for a range of cases.

* Clean up

* flake fixes

* Requested fixes

Also, added some tests and uncovered some corner cases.

Not insisting that, if the index is an sqlalchemy compountd construct,
it should be explicitly named using `.label()`. Raises exception
otherwise. This gets around trying to guess what pandas will name
that column.

* Further fixes

Split out npartitions and divisions to avoid confusion.

* cosmetics

* Trailing u

* Furhter small changes

* Make upper limit inclusive

Similifies what happens with max value

* Create SQL engine within task

Tests only worked on multi-threaded, because the sqlalchemy engine is not
serializable.

* flake

* Auto-guess chunksize if npartition not given

Also: enable operating on any input sqlalchemy expression, not only
a named table.

* more flake

* add read_sql_table to imports and docs

* Ensure time divisions are are range limits

12340 of 13341 relevant lines covered (92.5%)

0.92 hits per line

Source Files on job 5769.1 (PYTHON=2.7 NUMPY=1.10.4 PANDAS=0.19.0 COVERAGE='true' PARALLEL='false' TEST_IMPORTS='false' XTRATESTARGS=)