A SQLite vtable extension to read Parquet files
Go to file
Colin Dellow e2af2a07a4 Make rowid start from 1, not 0
Unclear whether this is strictly required, but I'm going to start using
SQLite as an oracle, and it'll be simpler if our rowids match theirs.
2018-03-18 17:03:46 -04:00
parquet Make rowid start from 1, not 0 2018-03-18 17:03:46 -04:00
parquet-generator Make rowid start from 1, not 0 2018-03-18 17:03:46 -04:00
tests Make rowid start from 1, not 0 2018-03-18 17:03:46 -04:00
.gitignore Generate queries from templates 2018-03-18 14:28:31 -04:00
LICENSE Initial commit 2018-03-02 18:37:08 -05:00
README.md Update README 2018-03-18 15:08:02 -04:00
build-sqlite Add script to fetch+build sqlite 2018-03-02 18:46:40 -05:00

README.md

sqlite-parquet-vtable

A SQLite virtual table extension to expose Parquet files as SQL tables.

Caveats

I'm not an experienced C/C++ programmer. This library is definitely not bombproof. It's good enough for my use case, and may be good enough for yours, too.

  • I don't use sqlite3_malloc and sqlite3_free for C++ objects
    • Maybe this doesn't matter, since portability isn't a goal
  • The C -> C++ interop definitely leaks some C++ exceptions
    • Obvious cases like file not found and unsupported Parquet types are OK
    • Low memory conditions aren't handled gracefully.

Building

  1. Install parquet-cpp
    1. Master appears to be broken for text row group stats; see https://github.com/cldellow/sqlite-parquet-vtable/issues/5 for which versions to use
  2. Run ./build-sqlite to fetch and build the SQLite dev bits
  3. Run ./parquet/make to build the module
    1. You will need to fixup the paths in this file to point at your local parquet-cpp folder.

Tests

Run:

tests/create-queries-from-templates
tests/test-all

Use

$ sqlite/sqlite3
sqlite> .load parquet/libparquet
sqlite> CREATE VIRTUAL TABLE demo USING parquet('parquet-generator/99-rows-1.parquet');
sqlite> SELECT * FROM demo;
...if all goes well, you'll see data here!...

Supported features

Row group filtering

Row group filtering is supported for strings and numerics so long as the SQLite type matches the Parquet type.

e.g. if you have a column foo that is an INT32, this query will skip row groups whose statistics prove that it does not contain relevant rows:

SELECT * FROM tbl WHERE foo = 123;

but this query will devolve to a table scan:

SELECT * FROM tbl WHERE foo = '123';

This is laziness on my part and could be fixed without too much effort.

Row filtering

For common constraints, the row is checked to see if it satisfies the query's constraints before returning control to SQLite's virtual machine. This minimizes the number of allocations performed when many rows are filtered out by the user's criteria.

Types

These Parquet types are supported:

  • INT96 timestamps (exposed as milliseconds since the epoch)
  • INT8/INT16/INT32/INT64
  • UTF8 strings
  • BOOLEAN
  • FLOAT
  • DOUBLE
  • Variable- and fixed-length byte arrays

These are not currently supported:

  • UINT8/UINT16/UINT32/UINT64
  • DECIMAL