On Wed, Feb 6, 2019 at 4:32 PM Pavan Deolasee <pavan.deola...@gmail.com> wrote:
>
> Hello,
>
> Currently either the table level option `toast_tuple_target` or the compile
> time default `TOAST_TUPLE_TARGET` is used to decide whether a new tuple
> should be compressed or not. While this works reasonably well for most
> situations, at times the user may not want to pay the overhead of toasting,
> yet take benefits of inline compression.
>
> I would like to propose a new table level option, compress_tuple_target,
> which can be set independently of toast_tuple_target, and is checked while
> deciding whether to compress the new tuple or not.
>
> For example,
>
> CREATE TABLE compresstest250 (a int, b text) WITH (compress_tuple_target =
> 250);
> CREATE TABLE compresstest2040 (a int, b text) WITH (compress_tuple_target =
> 2040);
>
> -- shouldn't get compressed nor toasted
> INSERT INTO compresstest250 VALUES (1, repeat('1234567890',20));
>
> -- should get compressed, but not toasted
> INSERT INTO compresstest250 VALUES (2, repeat('1234567890',30));
>
> -- shouldn't get compressed nor toasted
> INSERT INTO compresstest2040 VALUES (1, repeat('1234567890',20));
> INSERT INTO compresstest2040 VALUES (2, repeat('1234567890',30));
>
> Without this patch, the second INSERT will not compress the tuple since its
> length is less than the toast threshold. With the patch and after setting
> table level option, one can compress such tuples.
>
> The attached patch implements this idea.
>
I like this idea.
The patch seems to need update the part describing on-disk toast
storage in storage.sgml.
Regards,
--
Masahiko Sawada
NIPPON TELEGRAPH AND TELEPHONE CORPORATION
NTT Open Source Software Center
