About Reeza

Reeza · ‎06-27-2011

We'd need to see what the &mkt macro variable does resolve to to help out. If you're trying to create a format and have the data somewhere else consider using CNTLIN option instead of a macro variable. CNTLIN option allows you to create a format from a dataset.

Reeza · ‎06-24-2011

I think we need more info...if the data is online from your text, pointing to the link would be helpful and someone can duplicate your results. I noticed 2 variables named radiated...how many levels does each have. Are all Strata being tested at the same time in your text as well? Can always check the technotes for an errors in proc lifetest and/or your text for erratas...

Reeza · ‎06-24-2011

A couple of issues I can identify: In SAS you just subtract dates, if its date times then you need to take the date part, unless you want the time in seconds or milliseconds, not sure what the smallest unit of date time is. Also there is no CASE statement, there's a SELECT statement, but an if/then is probably what you're looking for. Your SQL statement is merging, but your datastep code is APPENDING the dataset not merging. I'm not 100% sure what you want but the following is a start with the following assumptions: date is in datetime you want difference in terms of days Merging files by client id and id *First need to sort both files: proc sort data=data_333; by client id; proc sort data=test_data; by client id; *then need to merge and get differences; Data Test; MERGE data_333 (in=a) test_data (in=b); by client id; *get only records from data_333; if a; if datepart(cd_date)-datepart(add_date))<=1 then cd_1=1; else cd_1=0; if datepart(cd_date)-datepart(add_date)<=2 then cd_2=1; else cd_2=0; run; OR SQL create table test as select a.add_date ,b.cd_date ,case when datepart(b.cd_date)-datepart(a.add_date) <= 1 then 1 else 0 end as cd_1 ,case when datepart(b.cd_date)-datepart(a.add_date) <= 1 then 1 else 0 end as cd_2 from data_333 a left join test_data b on a.client = b.client and a.id = b.id; quit;

Reeza · ‎06-24-2011

Rather than your initial Data _null_ step where you open the system via registry try opening by X command? You say it runs 3 times and then fails...are you looping the macro and thats where it fails or on the third try of the same open. You do have to make sure Excel has enough time to process the commands SAS sends to it and closes down other wise some funny things can happen. I'm also including a sample of what my close/quit looks like... x ' "C:\Program Files\Microsoft Office\Office11\excel.exe" '; filename cmdexcel; data _null_; x=sleep(5); run; data _null_; file cmdexcel; put &macrfile: run; .... other sas code data _null_; file cmdexcel; put '[save.as("new file name")]'; put '[close()]'; x=sleep(3); put '[quit()]'; run; HTH, Reeza

Reeza · ‎06-23-2011

Logged out and it does appear as my profile/nick name, not my full name. Interesting at least Pictures are not appearing properly however, ie IE7 or Firefox 3.6.6

Reeza · ‎06-23-2011

When posting in the new forum I'm seeing my full name. Do other see it or do they see my screen id and how can I tell. I assume there's a setting somewhere to check this?

Reeza · ‎06-23-2011

Does the program work if you replace the &macrfile with the actual file path not the macro variable? ie data _null_; file cmdexcel; put '[open("M:\ADC-P\CPSPB\Macros\Chdaily_test_.xlsm")]'; run; I typically use these options when using DDE: option noxwait noxsync mprint; the mprint option is really useful to help tell where the code is going wrong. Try posting the log (only) with the mprint option on as well.

Reeza · ‎06-22-2011

I think the dbtype= option might be what you want then. I don't know exactly wwere/how but hopefully google can help some more.

Reeza · ‎06-22-2011

That seems cumbersome. I suggest using proc import with type = mixed.

Reeza · ‎06-22-2011

You could expand the code to then include the name of the file in the where clause, ie where libname='SRC' and memname like 'indicators%'; Probably the easiest and most efficient solution. If you have multiple files with indicators_blah then you could restrict with the year or you could manually create the dataset the first time and automate the update procedure... create a macrovariable that contains the last day of the month and then update the dataset each time you have a new file. Then run your checks on that file, which should be the same every time I'm guessing. HTH, Reeza May need to tweak for case sensitivity. %let ddate=20110331; proc sql; insert into (dataset) select count(*), &ddate from src.indicator_&ddate; quit; OR proc sql; insert into (dataset) select memname, nobs nlobs from dictionary.tables where libname='SRC' and memname='indicators_&ddate"; quit;

Reeza · ‎06-21-2011

Do the old files change with each data dump? If not you can write a macro to just get the count from the latest file...if they do then you'll want to do it a different way. What type of files are you dealing with, ie text or SAS files? You may want to look at proc datasets/contents with the lib._all_ keyword. It also looks like your files are labelled with the last day of the month so could write a macro around that based on the date to loop through files. A simple way if all is sas, is to look at the dictionary.tables table proc sql; create table counts as select libname, memname, nobs, nlobs from dictionary.tables where libname='SRC'; quit;

Reeza · ‎06-20-2011

Thanks! I wanted 'banding' so I used a combination of methods, adding some things in my datastep and then in the proc report, the code provided definitely helped in getting there!

Reeza · ‎06-18-2011

Sorry, I don't think I was clear before, looking for varying the row colours based on group not the column colours. Thanks!

Reeza · ‎06-16-2011

Hi, I know I've seen this on here before but can't seem to find it at all. I'm using proc report to output some data and would like to alternate the background of the row colors based on the groups, so each group is differentiable from each other. I know I need to use a compute block and call define my style but not sure of the logic to get each group a different colour. Pointers in the right direction are appreciated. Thanks! Sample ID Value Colour 1 2 white 1 3 white 2 3 lightgrey 2 4 lightgrey 6 3 white 8 2 lightgrey Message was edited by: Reeza null

Reeza · ‎06-15-2011

Can you export with the libname statement or do you get an error? libname test excel "s:\EDPeds.xls" data test.ed; set edpeds; run; libname test;