apache / parquet-cpp / 425 / 1

DEFAULT BRANCH: master
Ran
Files 83
Run time 2s
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: 92.297% (-0.2%) from 92.51%
425.1

push

travis-ci

julienledem 
PARQUET-494: Implement DictionaryEncoder and test dictionary decoding

I incorporated quite a bit of code from Impala for this patch, but did a bunch of work to get everything working. In particular, I wasn't happy with the hash table implementation in `dict-encoder.h` and so have written a simple new one that we can benchmark and tune as necessary.

The simplest way to pull in the DictEncoder (PARQUET-493) was to also bring in the `MemPool` implementation, suitably trimmed down. We can continue to refactor this as needed for parquet-cpp.

I also did some light refactoring using `TYPED_TEST` in `plain-encoding-test` (now `encoding-test`).

Author: Wes McKinney <wesm@apache.org>

Closes #64 from wesm/PARQUET-494 and squashes the following commits:

c634abe [Wes McKinney] Refactor to create TestEncoderBase
a3a563a [Wes McKinney] Consolidate dictionary encoding code
2cc4ffe [Wes McKinney] Retrieve type_length() only once in PlainDecoder ctor
20ccd9e [Wes McKinney] Remove DictionaryEncoder shim layer for now
dcfc0aa [Wes McKinney] Remove redundant Int96 comparison
d98a2c0 [Wes McKinney] Dictionary encoding for booleans throws exception
05414f0 [Wes McKinney] Test dictionary encoding more types
9a5b1a4 [Wes McKinney] Enable include_order linting per PARQUET-539
f3f0efc [Wes McKinney] IWYU cleaning
d4191c6 [Wes McKinney] Add header installs, fix clang warning
1347b13 [Wes McKinney] Rename plain-encoding-test to encoding-test
09bf0fa [Wes McKinney] Fix bugs and add dictionary repeats
2e6af48 [Wes McKinney] Fix some bugs. FixedLenByteArray remains to get working.
69b5b69 [Wes McKinney] Refactor test fixtures to be less coupled to state. process on getting dict encoding working
6b23716 [Wes McKinney] Create reusable DataType structs for test fixtures and other compile-time type resolution matters
67883fd [Wes McKinney] Bunch of combined work for dict encoding support:

3331 of 3609 relevant lines covered (92.3%)

12427.62 hits per line

Source Files on job 425.1