SQL: how to generate "advanced" group by

I have to generate a group by using a function (e.g. Date()) like this

``` sql
Select column1, avg(column2) from table group by date(column1)
```

But how can I generate such a query using blaze? I tried the following:

```
from blaze import by
by(data.column1, col2=data.column2.mean())
```

But how to add the date-function?

UPDATE: I found a solution for pandas, but I would prefer doing it with blaze (the database): http://stackoverflow.com/questions/11391969/how-to-group-pandas-dataframe-entries-by-date-in-a-non-unique-column

Start at Blaze's `by` operation and inspect how the database expression is translated for grouping. Reproduce the requested `date(column1)` grouping against a supported database, then determine how completion should be verified from the generated query and result.

Written by the indexing model from the issue text.