Sum(Aggr(PD * Count(PD),PD))

[arethaking]

Hi All,

What is the meaning of this expression?

Sum(Aggr(PD * Count(PD),PD))

[varshavig12]

It will simple count PD and aggregate it with the same.

Example:

Here 1 has occurred twice , so its 1*2=2

and 2 has occurred once so its 2*1=2

# has occurred twice so, 3*2=6

and so on:

LOAD * INLINE [

PD

1

2

3

4

1

3

];

[varshavig12]

It will simply count the PD and multiply with its count.

[Kushal_Chawda · Accepted solution]

If you have sample data for this, you can create the straight table with Dimension & above expression and try to break the expression to understand how it's working

first see, what PD * Count(PD) is giving,

then see, what  Aggr(PD * Count(PD),PD) is giving

Finally see what is entire expression is giving

[sasiparupudi1]

aggr statement is similar to sql group by.. if you are familiar with sql.

PD * Count(PD) seems to be wrong..

select  PD,Count(Your Field)

group by (PD)

hth

Sasi

[jonathandienst]

This looks like a more complex (and less efficient) way to compute Sum(PD). For example, with this load script:

LOAD * INLINE

[

  PD

  .1

  .1

  .1

  .2

  .2

  .3

  .3

];

Sum(Aggr(PD * Count(PD), PD)) = 3 * .1 + 2 * .2 + 2 * .3 = 1.3

Sum(PD) = (.1 + .1 + .1) + (.2 + .2) + (.3 + .3) = 1.3