The compressed qual pushdown will look at presence of metadata columns
in the compressed chunk, not at whether the uncompressed column is an
orderby. This way, the decision on which columns have metadata is made
only during the creation of compressed chunk.

This doesn't introduce any behavior changes, just changes the decision
about which columns have metadata to happen in a single place. This will
simplify experimenting with configurable metadata columns.

Put global metadata columns first, segmentby next, orderby metadata
before the compressed orderby columns. This matches the order in which
they are likely to be accessed when reading from a compressed chunk.

…o. 7)

It is about 10x faster than row-by-row decompression, so if we're going
to decompress 10x less rows than the batch size, it's more efficient to
decompress row-by-row.

…r' and 'akuzm/rows' into HEAD

…o. 10)

…o. 11)

… 12)