About TimH

TimH · ‎09-19-2019

Ooops.. In my original code (not the 'pseudo code' I posted here) I had REALFLD1 and REALFLD2 reversed as they were close in spelling. Sorry for wasting everyone's brain time.

TimH · ‎09-19-2019

No, I want the two values to be in two separate fields, not concatenated. Please note the slash is '\' and not this '/' (which is one of the default delimiters for the function)

TimH · ‎09-19-2019

I have fields which contain either "String1" or they contain "String1\String2" with "\" as the delimiter between the two strings. I input the field with INPUT @col WORKFLD $20. I then have this code (following the example in the documentation for getting the first and last word in the string REALFLD1 = SCAN(WORKFLD,1,'\'); IF INDEX(WORKFLID,'\') > 0 THEN REALFLD2 = SCAN(WORKFLD,-1,'\'); It seems to work "backward" - when WORKFLD contains "OPSYS\KNMACTH" the results are REALFLD1 = KNMACTH and REALFLD2 = OPSYS . If I switch the '1' and the '-1' I get what I want but I'm not sure why it is working this way...

TimH · ‎09-05-2018

Thanks - the 'revised' example in that link will probably work for me. If it doesn't I will try TABULATE I guess.

TimH · ‎09-05-2018

Let me clarify - if I were to go to PROC MEANS or PROC TABULATE for this, I would need to create a variable for each of the plan names with a value of 1 or 0, and summarize all of those variables. That might work if we limit our selection to a finite group of plan names, and in this particular instance we might be able to do that. It would be better, though, to have the option to just have those values filled in on every line, even if they _are_ grouping variables.

TimH · ‎09-05-2018

Please note that I need the PLANNAME as an ACROSS variable.

TimH · ‎09-05-2018

A simple summary will not work because the "USE,PLANNAME" in the COLUMN definition in conjunction with the definition of PLANNAME as an ACROSS variable creates a column for each plan and this is what the requesting user desires. I proposed your solution but that is not what they want.

TimH · ‎09-05-2018

I need to write a CSV file from PROC REPORT output, but this CSV is intended to be loaded into a relational table. For that reason, I need each of the variables to be populated in every row of the CSV, whether they are grouping variables or not. The code is below. I have the OUT= because I was going to investigate the output to see if it was better to wrap the ODS code around a PROC PRINT of that dataset but I thought I'd ask here first. This code works fine, but the grouping variables in the CSV are in 'outline' form (the value is only populated when it changes) and that won't work for loading into a table. Is there something I can add to make them displayed on every row? ODS TAGSETS.CSVALL PATH="pathname" file="filenname.csv"; TITLE; PROC REPORT DATA=SUM2 OUT=MYRPT; BY WEEKBEG WEEKEND; COLUMN WEEKBEG WEEKEND DBNAME NAME AUTHID USES,PLANNAME; DEFINE WEEKBEG/ 'Week Beg' format=E8601DA10.; DEFINE WEEKEND/ 'Week End' format=E8601DA10.; DEFINE DBNAME/ 'Database' format=$8. group; DEFINE NAME/ 'Table' format=$32. group; DEFINE AUTHID/ 'ID' format=$10. group; DEFINE PLANNAME/ACROSS ' ' format=$11.; DEFINE USES/ANALYSIS ' ' SUM FORMAT=9.; COMPUTE BEFORE; DBNAME='*TOTAL*'; ENDCOMP; RBREAK BEFORE / summarize; RUN; ODS TAGSETS.CSVALL CLOSE; RUN;

TimH · ‎08-02-2018

Thanks - I found this right after I posted. I must not have been searching using the right terms. In any case, it works!

TimH · ‎08-02-2018

I have a dataset with variables DAYM1-DAYM31 for counts for each day of the month and DAYW1-DAYW7 for counts for each day of the week. It's easy to put column headers on these, but I would like to put a head _above_ the DAYM1-DAYM31 headers to say 'Day of Month' and above the others to say 'Day of Week' ... in PROC REPORT, how do I put that sort of heading? Do I need to create some sort of dummy variable?

TimH · ‎03-02-2018

We have the JDBC driver for MS SQL installed on z/OS (In a Unix Systems Services directory of course) and have used it successfully from within another query tool to access MS SQL data on z/OS. What can I do to enable using this JDBC driver in PROC SQL or a LIBNAME in the same way as we use DB2? It is entirely in Java and apparently uses no JNDI because we downloaded the .jar directly from Microsoft and didn't have to change anything for it to work. Seems like PROC SQL could use JDBC but I haven't had a lot of luck finding documentation on how to do it - searching for SAS+JDBC or "PROC SQL"+JDBC returns information about the SAS JDBC drivers which seem upon reading to provide JDBC access to SAS.

TimH · ‎02-02-2018

Ah, good old TABULATE. One of my favorite PROCs and that definitely will do what I want to do.

TimH · ‎02-02-2018

I am sure this was probably asked before but I don't know which keywords to bring up what I want - I searched and got a lot of PROC REPORT ACROSS variable stuff but nothing fits just yet, so I'm entering a new message. Here's what I want to do: I have datasets that have observations over time, with an identifier, date, and various measurement variables. I want to report, paged or otherwise grouped by identifier, with each date a column, and each variable a row with the variable label on the left and the values under the date columns. What is the best way to go: 1. Summarize (if necessary) the observations by date, then do PROC TRANSPOSE and PROC PRINT 2. PROC REPORT ? I am trying to figure out the latte but any and all advice is appreciated

TimH · ‎07-25-2016

Adding a reply here in case anyone else gets this as a recommendation. First - when transferring from z/OS to another platform: HTML and Javascript files (if any) should be transferred as Text (ASCII in some FTP programs) - codepage translaton will happen between EBCDIC and UTF8 or ASCII; image files (JPG, PNG, etc.) have to be transferred in IMAGE or BINARY mode. Secomd - there's a free Apache-based web server for z/OS that isn't hard to set up and get running; maybe just server the pages from the mainframe itself. One advantage: When you re-run your SAS program to tweak the graph or update it, you only need to refresh the web page in your browser, no file transfers necessary.

TimH · ‎07-19-2016

I think this is in the wrong area, but I can't find where to change it.

Online Status	Offline
Date Last Visited	‎03-21-2022 05:40 PM

Re: Problem with SCAN function?

Re: Problem with SCAN function?

Problem with SCAN function?

Re: PROC REPORT to CSV with values in all columns in each row

Re: PROC REPORT to CSV with values in all columns in each row

Re: PROC REPORT to CSV with values in all columns in each row

Re: PROC REPORT to CSV with values in all columns in each row

PROC REPORT to CSV with values in all columns in each row

Re: PROC REPORT - adding a spanning header over some columns

PROC REPORT - adding a spanning header over some columns

SAS z/OS support for IBM z15 INTEGRATED ACCELERATOR FOR Z SORT

Format Description at the Time of Definition

Re: Z/OS SMS Compression with Zedc

Z/OS SMS Compression with Zedc

Please provide JDBC support in SAS/Access

Add DESC= to DATA and PROC (and possibly ODS and others)

Re: Z/OS SMS Compression with Zedc

Re: Problem with SCAN function?

Re: Problem with SCAN function?

Problem with SCAN function?

Re: PROC REPORT to CSV with values in all columns in each row

Re: PROC REPORT to CSV with values in all columns in each row

Re: PROC REPORT to CSV with values in all columns in each row

Re: PROC REPORT to CSV with values in all columns in each row

PROC REPORT to CSV with values in all columns in each row

Re: PROC REPORT - adding a spanning header over some columns

PROC REPORT - adding a spanning header over some columns

Using JDBC with PROC SQL in z/OS batch or TSO

Re: PROC Report with groups across, statistics down?

PROC Report with groups across, statistics down?

Re: ODS output transfer from Z/Os

Re: SGPLOT on z/OS and HPGL plotter