Thread: The examples given in Chapter 14 are not true.
The following documentation comment has been logged on the website: Page: https://www.postgresql.org/docs/16/performance-tips.html Description: 1) How did you create the `tenk1` table? If you will claim that from the source code of version 9.3, the question arises how CREATE TABLE tenk1 (...) WITH OIDS; in v16? https://github.com/postgres/postgres/blob/524440579f838b30579f35d2045cc01b6f5070d8/src/test/regress/sql/create_table.sql#L37 Additionally, you have to spend time searching for additional and completely unnecessary information and old versions of the code, although you have just installed server version 16 from the source code, which has passed all checks - but as it turns out is not compatible with the documentation. After `make installcheck` v16 - All 215 tests passed. But 1) relpages for table `tenk1` - different in version 16. Some calculations will be incorrect accordingly 2) you will not be able to reproduce examples for which there are no indexes in table `tenk1` - for example index `index_tenk1_on_four`. 3) How do you see this familiarization with the product through documentation that points to older versions that work very differently in the new version?
PG Doc comments form <noreply@postgresql.org> writes: > Page: https://www.postgresql.org/docs/16/performance-tips.html > 1) How did you create the `tenk1` table? As the text says, it's supposed to be made by running the core regression tests ("make installcheck"). > 1) relpages for table `tenk1` - different in version 16. Some calculations > will be incorrect accordingly > 2) you will not be able to reproduce examples for which there are no indexes > in table `tenk1` - for example index `index_tenk1_on_four`. Yeah, that's a bit out of date. I fixed it up awhile ago for v17: https://git.postgresql.org/gitweb/?p=postgresql.git&a=commitdiff&h=5caa05749 However, the text also points out that you won't necessarily get exactly the same results as are shown, for assorted reasons of which different-version is only one. So I think these changes were more neatnik-ism than rescuing the chapter from total irrelevancy. regards, tom lane