Re: zheap: a new storage format for PostgreSQL

On Thu, Dec 6, 2018 at 11:13 AM Amit Kapila <amit.kapila16@gmail.com> wrote:
>
> On Thu, Dec 6, 2018 at 10:03 AM Pavel Stehule <pavel.stehule@gmail.com> wrote:
> >
> > čt 6. 12. 2018 v 5:02 odesílatel Mithun Cy <mithun.cy@enterprisedb.com> napsal:
> >>
> >> COPY command seems to have improved very slightly with zheap in both with size of wal and execution time. I also
didsome tests with insert statement where I could see some regression in zheap when compared to heap with respect to
executiontime. With further more investigation I will reply here. 
> >>
> >
> > 20% of size reduction looks like effect of fill factor.
> >
>
> I think it is because of smaller zheap tuple sizes.  Mithun can tell
> more about setup whether he has used different fillfactor or anything
> else which could lead to such a big difference.

Yes default fillfactor is unaltered, zheap tuples sizes are less and
alinged each at 2 Bytes

Length of each item. (all Items are identical)
=====================================
postgres=# SELECT lp_len FROM
zheap_page_items(get_raw_page('pgbench_zheap', 9)) limit 1;
 lp_len
--------
    102
(1 row)

postgres=# SELECT lp_len FROM
heap_page_items(get_raw_page('pgbench_heap', 9)) limit 1;
 lp_len
--------
    121
(1 row)

Total tuples per page
=====================================
postgres=# SELECT count(*) FROM
zheap_page_items(get_raw_page('pgbench_zheap', 9));
 count
-------
    76
(1 row)

postgres=# SELECT count(*) FROM
heap_page_items(get_raw_page('pgbench_heap', 9));
 count
-------
    61
(1 row)

because of this zheap takes less space as reported above.


--
Thanks and Regards
Mithun Chicklore Yogendra
EnterpriseDB: http://www.enterprisedb.com