Question

The following is a query that is taking too long; there are a few bits to it but it is otherwise a very simple query. I wanted to ask, can anyone see which part of it might be causing a performance bottleneck? Thanks a million.

SELECT DISTINCT node.nid AS nid
  , node_data_field_company_logo.field_company_logo_fid
    AS node_data_field_company_logo_field_company_logo_fid
  , node_data_field_company_logo.field_company_logo_list
    AS node_data_field_company_logo_field_company_logo_list
  , node_data_field_company_logo.field_company_logo_data
    AS node_data_field_company_logo_field_company_logo_data
  , node.type AS node_type, node.vid AS node_vid
  , node_data_field_company_logo.field_job_details_title_value
    AS node_data_field_company_logo_field_job_details_title_value
  , RAND() AS _random
FROM node node
  LEFT JOIN term_node term_node
    ON node.vid = term_node.vid
  LEFT JOIN term_data term_data
    ON term_node.tid = term_data.tid
  LEFT JOIN content_type_profile_job_post node_data_field_company_logo
    ON node.vid = node_data_field_company_logo.vid
WHERE (node.type in ('profile_job_post'))
  AND (node.status <> 0)
  AND ((term_data.name) = (''))
ORDER BY _random ASC

Answer 1

First, let’s get rid of all those useless aliases. And pay attention to the comments that asked for table structures, indexes, and EXPLAIN output.

SELECT DISTINCT node.nid AS nid
  , node_data_field_company_logo.field_company_logo_fid
  , node_data_field_company_logo.field_company_logo_list
  , node_data_field_company_logo.field_company_logo_data
  , node.type, node.vid
  , node_data_field_company_logo.field_job_details_title_value
  , RAND() AS _random
FROM node 
  LEFT JOIN term_node 
    ON node.vid = term_node.vid
  LEFT JOIN term_data 
    ON term_node.tid = term_data.tid
  LEFT JOIN content_type_profile_job_post 
    ON node.vid = node_data_field_company_logo.vid
WHERE (node.type in ('profile_job_post'))
  AND (node.status <> 0)
  AND ((term_data.name) = (''))
ORDER BY _random ASC

Simple things: you’ll usually benefit from an index on any column used in a JOIN clause or a WHERE clause.

node.status
node.type
node.vid
node_data_field_company_logo.vid
term_data.name
term_data.tid
term_node.vid

But it depends on the number of distinct values in each of those columns. For example, if node.status can have only two values, 0 and 1, then an index probably won’t help with it. (That’s one reason posting your table structures is important.)

Simple things revisited: you’re going to pay a heavy price for random order.

Slightly more complex things: you might benefit from creating a view on the node table. The view should select only the columns and rows you actually need here. Then JOIN on that view instead of using the node table directly.