sql - MySQL merge results into table from count of 2 other tables, matching ids -

i've got 3 tables: model, model_views, , model_views2. in effort have 1 column per row hold aggregated views, i've done migration make model this, new column views:

+---------------+---------------+------+-----+---------+----------------+ | field         | type          | null | key | default |          | +---------------+---------------+------+-----+---------+----------------+ | id            | int(11)       | no   | pri | null    | auto_increment | | user_id       | int(11)       | no   |     | null    |                | |  [...]        |               |      |     |         |                | | views         | int(20)       | yes  |     | 0       |                | +---------------+---------------+------+-----+---------+----------------+ 

this columns model_views , model_views2 like:

+------------+------------------+------+-----+---------+----------------+ | field      | type             | null | key | default |          | +------------+------------------+------+-----+---------+----------------+ | id         | int(11)          | no   | pri | null    | auto_increment | | user_id    | smallint(5)      | no   | mul | null    |                | | model_id   | smallint(5)      | no   | mul | null    |                | | time       | int(10) unsigned | no   |     | null    |                | | ip_address | varchar(16)      | no   | mul | null    |                | +------------+------------------+------+-----+---------+----------------+ 

model_views , model_views2 gargantuan, both totalling in tens of millions of rows each. each row representative of 1 view, , terrible mess performance. far, i've got mysql command fetch count of rows representing single views in both of these tables, sorted model_id added up:

select model_id, sum(c) (     select model_views.model_id, count(*) c model_views     group model_views.model_id     union     select model_views2.model_id, count(*) c model_views2     group model_views2.model_id) foo group model_id 

so nice big table following:

+----------+--------+ | model_id | sum(c) | +----------+--------+ | 1        | 1451   | | [...]    |        | +----------+--------+ 

what safest route pulling off commands here on in merge values of sum(c) column model.views, matched model.id model_ids out of above sql query? want fill rows models still exist - there model_views referring rows in model table have been deleted.

you can use update join on subquery:

update model    join (        select model_views.model_id, count(*) c         model_views        group model_views.model_id        union        select model_views2.model_id, count(*) c         model_views2        group model_views2.model_id) toupdate on model.id = toupdate.model_id set model.views = toupdate.c