ansaurus

Question

Help with Google App Engine query and datetime

Answer 1

+1 A:

In SQL you could do all sorts of fancy things, but Google API is rather limited.

Given that you want all records to be no more than 1 minute old, I'd just ask the database for all records less than 1 minute old, and then have python collate the results and reject the duplicate rows.

From the data you show here, it looks like you're getting a couple of rows per 'name' per minute or so, so that approach should be sufficient even though its inelegant.

The alternative would be to keep a second table with only the latest entry for each 'name' in it ... and cull that table every now and then to remove records over a minute old.

Sharkey 2009-04-16 01:31:34

Answer 2

+1 A:

I think I figured out a decent solution. The problem was in my model:

date = db.DateTimeProperty(auto_now_add=True)

This meant that for every instance of that model, the datetimes would all be slightly different. This makes grouping my data very difficult. So in my cron function, I made sure that every api request had the exact same time stamp.

The next change was to create a Current table. Every time the cron runs, it deletes everything in the Current table (only one row), and adds a new row. This new row is then added to a Log table which semi-permanently stores the results.

2009-04-16 15:28:02

Answer 3

+1 A:

Surely this would work:

query = db.GqlQuery("SELECT * FROM [table] ORDER BY date DESC LIMIT BY [num of rows]")

Alternatively, you could use an inequality, like "date > 2009-04-11 00:50", which would return all the results after that time.

pyguy 2009-06-20 22:08:16

ansaurus

tags:

views:

answers:

Help with Google App Engine query and datetime

related questions