About jcbell

jcbell · ‎08-26-2016

This may help someone... It took a bit of fiddling, but this works pretty slick. This uses DS2 with threads and a hash table. proc ds2; thread override / overwrite=yes; dcl double count; dcl bigint thisThread; dcl char(4096) classvartohash; dcl char(64) hashlookup; dcl package hash overridehash (); dcl package hiter hi('overridehash'); dcl char(128) scenario_name ; dcl char(128) scenario_type; dcl char(32) OVERRIDE_VARIABLE; dcl double precision OVERRIDE_VALUE_NUM; %do hashvars = 1;%to &numhashes; dcl char(64) hash_&hashvars ; %end; drop count; method init(); if 0 then set val.instvals (locktable=share); overridehash.definekey('hashlookup'); overridehash.data([SCENARIO_NAME SCENARIO_TYPE SCENARIO_DATE OVERRIDE_VARIABLE OVERRIDE_VALUE_NUM]); overridehash.dataset('{select * from work.recreate_overrides_upload_hash2 {options LOCKTABLE= SHARE} }'); overridehash.definedone(); end; method run(); dcl double rc; set val.instvals ; %do hashvars = 1 %to &numhashes; classvartohash=catt(&&classvartohash&hashvars); hash_&hashvars = SHA256HEX (classvartohash,0); hashlookup = hash_&hashvars; rc = overridehash.find(); if (rc = 0) then output; %end; thisThread=_threadid_; count+1; end; method term(); put '**Thread' _threadid_ 'processed' count 'rows:'; end; endthread; run; quit; proc ds2; data work.override_instid (keep=(instid scenario_name scenario_type hashlookup %do hashvars = 1 %to &numhashes; hash_&hashvars %end; SCENARIO_DATE OVERRIDE_VARIABLE OVERRIDE_VALUE_NUM) ); dcl thread override frac; method run(); set from frac threads=&override_threads; end; enddata; run; quit;

jcbell · ‎08-24-2016

Ok - Thanks. I think I mis-spoke. I believe you can use threads with a hash table. Peter Eberhardt talks about using a threaded DS2 step with a hash table in his book. The issue I'm running into is a locking of the lookup table by the threads. ERROR: BASE driver, Table RECREATE_OVERRIDES_UPLOAD_HASH is already in use and cannot be locked. If you opened the file with LOCKTABLE=EXCLUSIVE, try the LOCKTABLE=SHARE option I'm wondering if/how I can tell the hash object to open in Share mode. OR... Can I somehow reference _threadid_ and dynamically load a different hash table (work.foo__threadid_) for each thread. Perhaps using _NEW_ to create a new HASH instance for each thread? Curious. Thanks, John

jcbell · ‎08-24-2016

I have the same question - Anyone have an idea on this one. ?

jcbell · ‎08-01-2016

Just a quick thanks - The documentation is way off on this one and this was a big help. This may help someone else - This was my request file for Proc Soap. This STP has a couple of input parameters and a instream table. data _null_; file request; input; put _infile_; datalines4; <?xml version="1.0"?> <soapenv:Envelope xmlns:m="http://tempuri.org/stp_import_models" xmlns:soapenv="http://schemas.xmlsoap.org/soap/envelope/"> > <soapenv:Header> </soapenv:Header> <soapenv:Body> <m:basicWebService> <m:parameters> <m:_envusersk> 49 </m:_envusersk> <m:_getmodelnamesonly> NO </m:_getmodelnamesonly> </m:parameters> <m:streams> <m:instream> <m:Value> <Table> <InData> <uuid>a43f9000-0a14-0a6f-70de-c70aca70f6d9 </uuid> </InData> <InData> <uuid>0baecef9-0a14-0a6f-6ff1-c22da43b0bda</uuid> </InData> <InData> <uuid>07a45659-0a14-0a6f-6ff1-c22dc46c4391</uuid> </InData> </Table> </m:Value> </m:instream> </m:streams> </m:basicWebService> </soapenv:Body> </soapenv:Envelope> ;;;; run;

jcbell · ‎04-24-2016

I have a requirement where users would like to view models (SAS Code) and I'd like to use SAS Studio as the interface. Does anyone know of a way to pass Studio either the SAS code as a Object, stream on a macro variable or reference the code that is previously stored some place. I'd like to send users a URLthat when clicked opens Studio witht the code loaded in the editor. Thanks, John

jcbell · ‎03-13-2014

Hello, I'm having a heck of a time getting an absolute layout to consume the entire physical page. Despite my attempts, the region thinks it's too small for the content and states that output is lost. There is still room on the physical page for output, but layout complains and quits writing to the region. I have a region statement: ods region x=0in y=2.791in width=10in height=30in ; ... then 2 proc report statements The dataset has 48 obs and is printing 33 of them. I can adjust the report to squeeze things into the space - but I can't seem to get the region to extend to the bottom of the page. I have margins set to zero and they seem to be working fine because my top margin is right at the top of the page. options leftmargin=0 in rightmargin= 0 in bottommargin= 0 in; Any ideas or any ideas how to see the region to help debug? I suspect the region is not as large as I think it is. Thanks, John

jcbell · ‎11-03-2013

Cynthia, Thanks - I'm afraid you are right. I tried using an image as well - John

jcbell · ‎11-01-2013

Cynthia, As always - thanks for your time. Looking back - a bit more detail is probably helpful. I've attached a self-contained data/code snipet that reproduces the challenge. The attached image with the yellow highlighting is the area of interest. I've tried both INDENT and MARGIN as ways of indenting the text. What I can't seem to figure out is a clean way of indenting the background color so that the entire line is not highlighted. I've using PDF as the output for ODS. Let me know what you think. Thanks, John

jcbell · ‎11-01-2013

Hello, Curious if anyone can think of a creative way to pull off this requirement: The requirement sounds simple - We have some lines of the report that require indenting and highlighting. The problem is that the client would like the highlighting to also be indented. Another words, we don't want the entire row highlighted. See attachment. The indenting is working fine and the highlights are working fine - but the background hightlighting affects the entire row of data (see proc_print_indent2). I'm using a fairly straightforward compute statement: call define(_col_,"Style", "Style=[foreground=black indent=20 just=l background = &_rpt_bkgrnd_2 BORDERLEFTWIDTH=1 BORDERLEFTCOLOR=black BORDERBOTTOMCOLOR=black BORDERBOTTOMWIDTH=1 BORDERTOPWIDTH=1 BORDERTOPCOLOR=black ]"); Any ideas on how I could fill-in the background to match the indent? Thanks, John ps - this is the curse of Excel...

jcbell · ‎06-12-2013

Simply solution - I didn't realize that streaming was set on the STP. Simply set to package and the output from sgrender works nicely. Thanks, John

jcbell · ‎06-12-2013

Hello, Does any one have a good methodology for streaming sgrender output in a stored procedure to Excel via the SAS add-in? I have a stored procedure that users access via 'Reports' under the MS add-in. The STP is running sgrender and I'd like to stream or at least display the output in Excel. Thanks, John

jcbell · ‎03-26-2012

The line you have will connect you to the metadata server just like JMP. Accessing files from Base SAS would require a drive to be mapped to your compute tier or filesever where your programs reside. If you have access to EG you can use EG to access files on a remote system.

jcbell · ‎03-24-2012

I think you may need a couple of steps: Something like this should work: proc sql; create table work.test (query_for_test1 char(10),... etc.. rundate NUM format=datetime16.); insert into work.test (rundate) values(%sysfunc(datetime())); insert into work.test from work.query_for_test1.... etc. (real data) quit;

jcbell · ‎03-24-2012

The IN statements allow you to do a match merge. Using combinations of the IN statements you can replicate LEFT, RIGHT and FULL join logic that is found in SQL. In your statement above, a or b =1 if a matching record is found. if a=1 or b=0: Record found in A and Not in B a=1 and b=1: Full Join - records found in each table b=0 or b=0: No record found in B - not sure why you would merge a=0 and b=0: Records not found in either Also from @Tom: Here is a paper that talks about SAS merging. That might help. http://www2.sas.com/proceedings/sugi22/ADVTUTOR/PAPER39.PDF The IN= variables are indications of whether the dataset contributed information to the current merge. You can think of it as an indication of whether there are any observations in that source dataset for this BY variable grouping. For left or right join you just need to reference A or B , respectively. Whether the value was in the other datasets doesn't matter. The big difference between merge and SQL join is when both datasets have multiple observations in the same BY variable group. SQL JOIN will generate all of the combinations. The merge statement will just match the records in the order that they come. If one dataset runs out of records to contribute for that BY variable grouping then the values from the last observation read stay unmodified and are matched to the rest of the records from the other dataset.

jcbell · ‎03-24-2012

Depending upon what type of terminology you are searching for, there are several potential sites to review: http://www.arena-international.com/ebf/ http://www.bankersonline.com/forum/ubbthreads.php?ubb=cfrm&c=2 http://www.topix.com/forum/business/investment-banking good site for finance/risk: http://www.garp.org/

Online Status	Offline
Date Last Visited	‎01-04-2022 10:51 AM

Re: Can DS2 threads share a single hash object?

Re: Can DS2 threads share a single hash object?

Re: Can DS2 threads share a single hash object?

Re: Help with XML Syntax for post request to Stored Process

Automatically open SAS Code

ODS Absolute layout proc report

Re: Proc Report, Indent and background color

Re: Proc Report, Indent and background color

Proc Report, Indent and background color

Re: Add-in MS Office stream sgrender ODS output

Re: Automatically open SAS Code

Automatically open SAS Code

Re: Can DS2 threads share a single hash object?

Re: Can DS2 threads share a single hash object?

Re: Can DS2 threads share a single hash object?

Re: Help with XML Syntax for post request to Stored Process

Automatically open SAS Code

ODS Absolute layout proc report

Re: Proc Report, Indent and background color

Re: Proc Report, Indent and background color

Proc Report, Indent and background color

Re: Add-in MS Office stream sgrender ODS output

Add-in MS Office stream sgrender ODS output

Connect to Metadata server from Base SAS

No results still populate a field?

IN Dataset option

Banking Discussion Forum