P() and E()

[m_perreault]

Hi All,

Was wondering if anyone knows if using P() or E() is faster holding all else equal.  Basically I have the option use an expression sum({<[ID] = E({<[Include Flag] = {'0'}>}ID)>}Sales) or sum({<[ID] = P({<[Include Flag] = {'1'}>}ID)>}Sales) and am wondering if one is better than the other.

Thanks,

Mark

[tresesco]

I have never tested to compare their performances. However, I guess if there is a winner, then it should be the p() because that is explicit. 

[jonathandienst]

YMMV - you need to test this with your particular data set as that may influence which expression is quicker (if any).

[marcus_sommer]

Like the others I never made systematically testings to compare the performance of p() and e(). But if I would do a numerical check and using 0 or 1 instead of '0' or '1'.

Beside this I would also try if the flag couldn't be used more directly to filter/calculate the right data. I mean something like:

sum(Sales * [Include Flag])

- Marcus