Question

I have a table that contains the following columns:

ID int, DISTANCE float, EVENT Varchar

What i would like to acheive is to select all the data, but to group by the event and distance fields (i.e. to remove duplicate events at the same distance).

The issue I can see is that the Distance column is a float, and as such the GROUP BY may not behave expectantly. Most (i say most) of the data is stored to 3 decimal places, and this is the criteria I would like to group on.

Example Data:

ID,Distance,Event

1, 0.001, A
2, 0.002, A
3, 0.002, A
4, 0.002, B
5, 0.003, C
6, 0.0035,C

So the result would look like:

1, 0.001, A
2, 0.002, A
4, 0.002, B
5, 0.003, C

Answer 1

You can use CAST() to convert the decimal to string and only use the first 5 characters of the converted string ("x.xxx"), which you can then group on with GROUP BY:

Ex:

SELECT
    ID,
    CAST(Distance AS CHAR(5)) AS DistanceGroup,
    Event
FROM 
    yourtbl
GROUP BY
    DistanceGroup,
    Event

EDIT: You can also use TRUNCATE() on the Distance field which may actually be better in terms of performance:

SELECT
    ID,
    TRUNCATE(Distance, 3) AS DistanceGroup,
    Event
FROM 
    yourtbl
GROUP BY
    DistanceGroup,
    Event

Additional info on CAST() and TRUNCATE()