tree e8d387660ee251d3d750491614c2606602982a9a
parent 8bd769bc118f617580cfaa7b1e654159c0fb696b
author Marco Neumann <marco@crepererum.net> 1620665098 +0200
committer GitHub <noreply@github.com> 1620665098 +0200
gpgsig -----BEGIN PGP SIGNATURE-----
 
 wsBcBAABCAAQBQJgmWMKCRBK7hj4Ov3rIwAAdKAIAAJAAaqoSFv8PkRCgAX+X1tU
 ryUx8umSvnU+jb98CEDHBFOdqAYr6ja0vuKhusSUe7I9VRlevgIq8RI6XuEPG7ca
 SNWn7PteAQf05FpLx7Tau5ztPz0WmE6Ek7Ib6ALnpKKkf6/IrXwDfsdLCumT/anW
 wrwninlDB2/xsGJwr9rjFb7dNgp+967x06FCyI7Tne8LRe42xd/s1F1mKSy2rbpv
 u/KHxKx6LFKqm3dRtY1RHYJITJwkzjzC2wGaRLjSC8JncJ1dmBaXoogglbV2ne2r
 c53YWZvIeeN/5SRTZCiQS1HNYDLn52zhTea7oum+5g53zZ5AhYtB2UwYySZBG+U=
 =6/EO
 -----END PGP SIGNATURE-----
 

support full u32 and u64 roundtrip through parquet (#258)

* re-export arity kernels in `arrow::compute`

Seems logical since all other kernels are re-exported as well under this
flat hierarchy.

* return file from `parquet::arrow::arrow_writer::tests::[one_column]_roundtrip`

* support full arrow u64 through parquet

- updates arrow to parquet type mapping to use reinterpret/overflow cast
  for u64<->i64 similar to what the C++ stack does
- changes statistics calculation to account for the fact that u64 should
  be compared unsigned (as per spec)

Fixes #254.

* avoid copying array when reading u64 from parquet

* support full arrow u32 through parquet

This is idential to the solution we now have for u64.