From: Reindl Harald
Date: September 27 2012 9:25pm
Subject: Re: checking progress of alter table on a MyISAM table
List-Archive: http://lists.mysql.com/mysql/228281
Message-Id: <5064C45B.8020804@thelounge.net>
MIME-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha1;
 protocol="application/pgp-signature";
 boundary="------------enig89E9852AE44F11E9F1A692AD"

--------------enig89E9852AE44F11E9F1A692AD
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

Am 27.09.2012 23:15, schrieb Larry Martell:
>> Indexes are slowing this down. your calculation of 79 hours should be
>> correct, only if there are no unique indexes, otherwise this will slow=
 down
>> more as the data increases.
>=20
> It is what it is - there's no free lunch - dropping the indexes, doing
> the alter, and re-adding the indexes would probably take just as long

you refuse to understand the implication of indexes
it would NOT take so long - this was proven thousands
of times by many people all over the world

having a unique index on a table with many rows means
a lot of overhead for EVERY single insert/update

without the key you avoid the complete overhead
and do it once at the end - why do you not simplay
make any test table and compare insert 50.000 rows
with and without keys on the test-table

this takes a few minues and after that you understand
how large the overhead is and that it is grwoing with
the table


--------------enig89E9852AE44F11E9F1A692AD
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: OpenPGP digital signature
Content-Disposition: attachment; filename="signature.asc"

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://www.enigmail.net/

iEYEARECAAYFAlBkxFsACgkQhmBjz394AnkFjACfafENMOw/nQuW5VCv4PttUpi3
jlkAoInKcgE7v49ywr+i9cMjZDH00vZ4
=B7sD
-----END PGP SIGNATURE-----

--------------enig89E9852AE44F11E9F1A692AD--