On Thu, Sep 22, 2016 at 1:37 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> OK, I'll think about how to do that more efficiently. The smaller
> incremental improvement isn't surprising, because in this example the
> index would still be 90-something MB if it had no free space at all,
> so there's going to be decreasing returns from any additional work
> to avoid wasted free space. But if we can do it cheaply, this does
> suggest that using pages in order by free space is of value.
Tom, are you planning to do something about this patch yet this
CommitFest, or leave it until later?
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company