Re: newbie: how to return one iteration *per unique date (DAY!)* in a timestamp column?

Chris wrote:

Ben Dunlap wrote: ...which will echo:

"<tr><td>#records in 't7solar_landing' matching the given (iterating) date (in the 'solarLandingDateTime' column)</td><td>#records in 'aw_7solar_confirm' matching the given (iterating) date (in the 'solarAwConfDateTime' column)</td><td>#records in 'aw_7solar_aw' matching the given (iterating) date (in the 'solarAWDateTime' column)</td></tr>"...

If you just need to count the records with a particular date you should be able to use this construction: SELECT COUNT(*) AS record_count,

       DATE(<date_column>) AS date_field

FROM <table> GROUP BY date_field You could probably write a generalized PHP function (called 'build_query()' or something) that would construct this query given a table name and a date-column name, and call it once for each table/column pair. Then you could stitch the three query strings together, in PHP, into one large query using SQL's "UNION ALL", which concatenates the results of multiple queries into one large result-set: (<query 1>) UNION ALL (<query 2>) UNION ALL (<query 3>) And then pass that one large query to the database.

.. and a field describing which table it came from, otherwise you end up with: count | date ------------

5     | 2009-01-01
10    | 2009-01-01

and no reference point.

I should have given an example .. select count(*) as record_count, date(column_name) as date_field, 'my_table' as table_name union all select count(*) as record_count, date(column_name) as date_field, 'my_table_2' as table_name and end up with:

count | date       | table_name

-------------------------------

5     | 2009-01-01 | table 1
10    | 2009-01-01 | table 2

-- Postgresql & php tutorials http://www.designmagick.com/