BookmarkSubscribeRSS Feed
zion1
Calcite | Level 5

Hi, relatively new SAS user here. Any help is greatly appreciated!

 

I need a table with the number of occurrences in each month by 2 criteria (TP and Elevated). So for the following example, I need a table with the # of TP=1 and Elevated=1 occurring in January.

 

example data

ID  DOV               Elevated  TP

1   01JAN2017     1               1

2   05JAN2017     0               2

3   07JAN2017     1               3

4   01MAR2017    0               1

5   09MAR2017    1               1

 

I am able to calculate number of occurrences in each month by TP using:

proc sql;

CREATE TABLE screen_mth_detail as

SELECT distinct(TP), COUNT(distinct ID)AS TP_Count,

SUM (DOC between '01AUG2017'd AND '31AUG2017'd)AS AUG17_Count,

SUM (DOC between '01SEP2017'd AND '30SEP2017'd)AS SEP17_Count,

SUM (DOC between '01OCT2017'd AND '31OCT2017'd)AS OCT17_Count,

SUM (DOC between '01NOV2017'd AND '30NOV2017'd)AS NOV17_Count,

SUM (DOC between '01DEC2017'd AND '31DEC2017'd)AS DEC17_Count

FROM epds1

Group BY TP;

run;

 

How can I calculate the same but for only the cases in which Elevated=1?

11 REPLIES 11
novinosrin
Tourmaline | Level 20
I need a sample HAVE and sample WANT( your required output) for me to understand as I am dumb unlike others. Requesting you for the same with a brief definition of logic. Thank you
Reeza
Super User

Please supply more example data that illustrates why you need DISTINCT. 

Nothing shown indicates the need for DISTINCT and if you don't need it, it's much easier to switch to a SAS PROC such as PROC MEANS which will aggregate data by month automatically without you having to hardcode the time intervals. 

zion1
Calcite | Level 5

Looks like I may be making this more complicated than it needs to be.  My data looks like this:

IDDOVAgedScoreElevateTP
138-Aug-1732301
119-Aug-1711951413
199-Aug-1743002
149-Aug-17421412
1015-Aug-172751413
1415-Aug-17191213
1716-Sep-1735002
1817-Sep-17681412
1517-Sep-1732301

 

I need an output like this: (count of visits that are at TP1-3 in each month and of those, how many were elevated (elevate=1)

 TP 1TP 1 ElevatedTP 2TP 2 ElevateTP 3TP 3 Elevated
August102133
Sept102100

 

Thank you for your reply!

PGStats
Opal | Level 21

Try this:

 

data have;
input ID DOV :anydtdte. Aged Score Elevate TP ;
datalines;
13 8-Aug-17 32 3 0 1 
11 9-Aug-17 1195 14 1 3 
19 9-Aug-17 43 0 0 2 
14 9-Aug-17 42 14 1 2 
10 15-Aug-17 275 14 1 3 
14 15-Aug-17 19 12 1 3 
17 16-Sep-17 35 0 0 2 
18 17-Sep-17 68 14 1 2 
15 17-Sep-17 32 3 0 1 
;

proc freq data=have;
format DOV monname10.;
table DOV*tp*elevate / out=freqs sparse noprint;
run;

data fixed;
set freqs;
id = catx(" ", "TP", tp, ifc(elevate, "Elevated", ""));
run;
 
proc transpose data=fixed out=want(drop=_: );
by DOV;
id id;
idlabel id;
var count;
run;
PG
pau13rown
Lapis Lazuli | Level 10

just create a var using the month function ie month=month(), and then use proc freq as someone else indicated, except use the 'list' options on the table statement ie tables month*tp*elevate /list;

Reeza
Super User

@pau13rown wrote:

just create a var using the month function ie month=month(), and then use proc freq as someone else indicated, except use the 'list' options on the table statement ie

tables date*tp*elevate /list;
format date monname3.;

;


That's not required, in SAS you can apply the month format within PROC FREQ so you don't need a new variable. 

 

 

pau13rown
Lapis Lazuli | Level 10

'not required' depends on your temperament.Having a variable that is merely implied doesn't encourage defensive coding etc

Reeza
Super User

What's 'defensive coding'?

pau13rown
Lapis Lazuli | Level 10

it's a euphemism for paranoia in the drug industry. For example, if i "know" my text variable (lab parameter for example) is in caps, i am still writing my code as "if upcase(param)='HCT' then ...". rather than simply "if param='HCT' then ..." If you make an error in industry there is nothing you can say to get you out of it, the cost is too high, hence defensive coding

PGStats
Opal | Level 21

I agree with the virtues of defensive coding. I consider the temporary association of a format Inside an analysis procedure such as proc freq to be defensive against any format association that might have been made outside the procedure. It also défends against the multiplication of versions of your data.

PG
pau13rown
Lapis Lazuli | Level 10

well, we could waste time discussing it, the industry is a different universe and we'll speak past each other. There would be no multiplication of data, obviously. The analysis plan would stipulate or imply that the variable is needed, thus it would exist in a permanent dataset, refer to cdisc, sdtm, adam. The format applied would be indicated within that documentation etc. When i said 'etc' above i was alluding to validation, you cannot validate a variable that doesn't exist. The process is extremely pedantic, maybe they document things with proc compare etc. Thus, throwing a format within freq shows an old school nonchalance that doesn't exist anymore. Each to his own, but i would try to encourage (only incidentally) good programming practice when discussing code: http://www.phusewiki.org/wiki/index.php?title=Good_Programming_Practice

sas-innovate-2024.png

Join us for SAS Innovate April 16-19 at the Aria in Las Vegas. Bring the team and save big with our group pricing for a limited time only.

Pre-conference courses and tutorials are filling up fast and are always a sellout. Register today to reserve your seat.

 

Register now!

How to Concatenate Values

Learn how use the CAT functions in SAS to join values from multiple variables into a single value.

Find more tutorials on the SAS Users YouTube channel.

Click image to register for webinarClick image to register for webinar

Classroom Training Available!

Select SAS Training centers are offering in-person courses. View upcoming courses for:

View all other training opportunities.

Discussion stats
  • 11 replies
  • 1839 views
  • 1 like
  • 5 in conversation