topic Re: Proc report or tabulate for multiple groups with single row entry for analysis in ODS and Base Reporting

Proc report or tabulate for multiple groups with single row entry for analysis

Swapnil_21 — Sat, 28 Sep 2019 02:50:54 GMT

Hi All,

I have 4 groups in my data and I need to summarize based on each column but no repetition of rows. Data looks like this

Zone	Region	ZH_Abbreviation	RH_Abbreviation	Expense_In_Thousands
Western	Maharashtra	MH	Rahul	100
Western	Maharashtra	MH	Ajay	600
Western	Maharashtra	MH	Ajay	200
Western	Maharashtra	NGP	Rahul	400
Western	Maharashtra	NGP	Pankaj	450
Western	Gujarat	AM	Ashok	600
Western	Gujarat	AM	Ajay	400
Western	Gujarat	GK	Pankaj	300
Eastern	Bengal	AM	Pavan	300
Eastern	Bengal	AM	Rajesh	400
Eastern	Tripura	GK	Rohan	300

And I need output like below

Zone	Region	ZH_Abbreviation	RH_Abbreviation	Expense_In_Thousands
Western	Maharashtra	MH	Rahul	100
			Total for Rahul	100
			Ajay	800
			Total for Ajay	800
		Total for MH		900
		NGP	Rahul	400
			Total for Rahul	400
			Pankaj	450
			Total for Pankaj	450
		Total for NGP		850
	Total for Maharashtra		1750
Total for Western			1750
Eastern	Bengal	AM	Pavan	300
			Total for Pavan	300
			Rajesh	400
			Total for Rajesh	400
		Total for AM		700
	Tripura	GK	Rohan	300
			Total for Rohan	300
		Total for GK		300
	Total for Tripura		300
Total for Eastern			300
Grand Total				2050

I tried with the Proc report and Group by on 4 columns but it results in repetition of cell values. Is this doable ? Any help is really appreciated.

Thanks in advance

Swapnil

Re: Proc report or tabulate for multiple groups with single row entry for analysis

SASKiwi — Sat, 28 Sep 2019 06:23:51 GMT

Absolutely doable in PROC REPORT. Define ZONE, REGION, ZH_Abbreviation, RH_Abbreviation as GROUP columns, and Expense_In_Thousands as an ANALYSIS column.

Then add BREAK AFTER <Varname> / SUMMARIZE statements for ZONE, REGION, ZH_Abbreviation, RH_Abbreviation.

Lastly add a RBREAK AFTER / SUMMARIZE statement.

Re: Proc report or tabulate for multiple groups with single row entry for analysis

Swapnil_21 — Sat, 28 Sep 2019 08:31:34 GMT

@SASKiwi I have tried that way. However, result is not as per our expectations. It shows repeated values on the groups.

Proc report data=have;

column zone Region ZH_Abbreviation RH_Abbreviation Expense_in_Thousands;

define Zone / Group "Zone"

define ZH_Abbreviation / Group "ZH Abbreviaton";

define RH_Abbreviation / Group "RH Abbreviation";

define Expense_In_Thousands/ Analysis;

break after Zone / summarize;

break after Region / summarize;

break after ZH_Abbreviation / summarize;

break after RH_Abbreviation / summarize;

RBREAK AFTER/ Summarize;

Run;

It produces multiple rows per group. However, I want only first value of the member and rest to be blank.

Re: Proc report or tabulate for multiple groups with single row entry for analysis

Cynthia_sas — Sat, 28 Sep 2019 15:27:50 GMT

What happened to Gujarat and AM under Western? What order are the rows supposed to be in? Normally the G in Gujarat would sort before the M in Maharashtra. And, Ajay would sort before Rahul. Just curious because what you show here seems different than what you would get.
The break variable information will be repeated unless you use SUPPRESS on the Break statement. However, if you want to customize the text at the break, you'll need a COMPUTE block.

But I wouldn't bother with using SUPPRESS, since you need the values at the break in order to customize the BREAK line as shown below for HTML output:

Basically a COMPUTE block for an AFTER location can "touch" all the other cells on the row, so in the COMPUTE blocks below, after I customize the main cell for the COMPUTE block, I blank out the other cells on the row:

That did require a LENGTH statement when the variables were created so that the cells would be wide enough for the customized text:

I added order=data to the PROC REPORT DEFINE statements to keep the report in the same order as the original data. Otherwise everything would be in alpha order.

Cynthia