Easier way to do this w/out boggin Excel?

[PCRIDE]

Hi, I build a lot of dashboards with just raw data, my data sets are calculating different metrics for different groups.

I have always done Sumifs and created the metrics by criteria. The problem with this is some of the metrics formulas can be quite large and before I know it my formula is 6 rows long. I have noticed this starts to bog down excel. I also create addtional helper cells to get 1 or 0 to make some of the formulas shorter.

The other problem I have is when there is no data or dividing by 0 I get the #DIV/0!, is there any other way of making this blank other than using ISERR or any of those? Because using those you have to double up the formula and it becomes very large, If(ISERR ,(6 nested formulas do this, otherwise do that.

Here is my formula.

=SUMIFS(TT_TT!$EI:$EI,TT_TT!$DY:$DY,SD!$S$52,TT_TT!$DP:$DP,SD!O58)/(SUMIFS(IM_RF!$EG:$EG,TT_TT!$EI:$EI,"0",TT_TT!$DY:$DY,SD!$S$52,TT_TT!$DP:$DP,SD!O58)+SUMIFS(TT_TT!$EI:$EI,TT_TT!$DY:$DY,SD!$S$52,TT_TT!$DP:$DP,SD!O58))

If I wanted to avoid the DIV/0 error the formula would be doubled.

=if(iserrr(SUMIFS(TT_TT!$EI:$EI,TT_TT!$DY:$DY,SD!$S$52,TT_TT!$DP:$DP,SD!O58)/(SUMIFS(IM_RF!$EG:$EG,TT_TT!$EI:$EI,"0",TT_TT!$DY:$DY,SD!$S$52,TT_TT!$DP:$DP,SD!O58)+SUMIFS(TT_TT!$EI:$EI,TT_TT!$DY:$DY,SD!$S$52,TT_TT!$DP:$DP,SD!O58),"--",SUMIFS(TT_TT!$EI:$EI,TT_TT!$DY:$DY,SD!$S$52,TT_TT!$DP:$DP,SD!O58)/(SUMIFS(IM_RF!$EG:$EG,TT_TT!$EI:$EI,"0",TT_TT!$DY:$DY,SD!$S$52,TT_TT!$DP:$DP,SD!O58)+SUMIFS(TT_TT!$EI:$EI,TT_TT!$DY:$DY,SD!$S$52,TT_TT!$DP:$DP,SD!O58))

 

[PCRIDE]

My data sets are about 50K rows and sometimes larger, up to 100K

 

[Ganjin]

hmm, if you had excel 2007 or higher, you could use:

=IFERROR(SUMIFS(TT_TT!$EI:$EI,TT_TT!$DY:$DY,SD!$S$52,TT_TT!$DP:$DP,SD!O58)/(SUMIFS(IM_RF!$EG:$EG,TT_TT!$EI:$EI,"0",TT_TT!$DY:$DY,SD!$S$52,TT_TT!$DP:$DP,SD!O58)+SUMIFS(TT_TT!$EI:$EI,TT_TT!$DY:$DY,SD!$S$52,TT_TT!$DP:$DP,SD!O58)),"--")

as for the sumifs ...
I would be inclined to suggest sumproduct, but since you will have to convert something like A:A to A1:A1000, that might wind up making your formula larger ...

anywho, good luck

 

[PCRIDE]

great! I do have 2010 so this will work. Thanks just never noticed that one...