Re: Leftover reference to replacement selection 1 run case - Mailing list pgsql-hackers

From Peter Geoghegan
Subject Re: Leftover reference to replacement selection 1 run case
Date
Msg-id CAH2-WzkmR5PCexYoXRq_s9xfxj5x4+E5vQdqnuh3E1DsObMvww@mail.gmail.com
Whole thread Raw
In response to Re: Leftover reference to replacement selection 1 run case  (Robert Haas <robertmhaas@gmail.com>)
Responses Re: Leftover reference to replacement selection 1 run case
List pgsql-hackers
Yes, you have that right. Per dumptuples(), even the zero tuple run edge case will still write a run marker, and will therefore still consume a tape. We must have at least two initial runs to merge. (though dummy runs for non final merges are a slightly different matter.) 

--
Peter Geoghegan
(Sent from my phone)

pgsql-hackers by date:

Previous
From: David Steele
Date:
Subject: Re: PATCH: Exclude unlogged tables from base backups
Next
From: Michael Paquier
Date:
Subject: Re: money type's overflow handling is woefully incomplete