I use PC SAS (Display Manager) 9.4 (TS1M3) to connect to a Linux server with the same version of SAS.  I have a new user to whom I forwarded my SAS/Connect program + the tcp script.  When he runs it using his ID and password, he gets this error:   NOTE: Remote signon to P133 commencing (SAS Release 9.04.01M3P062415). NOTE: Script file 'tcp_p133_ssh.txt' entered. ERROR: Script file interpretation terminated due to error. ERROR: A communication subsystem partner link setup request failure has occurred. ERROR: Read error. ERROR: Remote signon to P133 canceled.   And... in this New Message window, I don't know what "Labels" are vs "Message Tags". ... View more

One of my quant users is getting this message on a Linux 6/SAS 9.4 (TS1M3) Proc NLMIXED job:  proc nlmixed data=testda_ft maxiter=800 maxfunc=8000 nthreads=24;   and she gets this     "WARNING:  The NTHREADS= option is ignored due to insufficient memory."   She ran it from a Putty command line with this:  /usr/local/SAS/SASFoundation/9.4/sas -work /lppra00a0133/temp/saswork -noterminal scoring_GN30.sas &   It loops month by month through the data and I asked her to run 1 month with MPRINT and FULLSTIMER turned on. Log is attached.     /*tuning for one category product */ %macro mbspred(product_name, inputsasdb , month_reset_min, month_reset_max , resetperi_min, resetperi_max , orig_term_min, orig_term_max , wala_min, wala_max , obsdt_bgn, obsdt_end , vintage_begin, vintage_end , agency_type, product_type, coll_type , exclude_mon , Parm_inputda_dir, tunParmInFile , nlinout_db , output_dir , nlin_graph_nm, tunParmOutFile, tunSumOutFile , model_num, nlin_sum ); %dafilter_ARM( &inputsasdb.,outab=testda_ft ,mth2reset_min=&month_reset_min. , mth2reset_max=&month_reset_max. ,resetperi_min=&resetperi_min. , resetperi_max=&resetperi_max. ,orig_term_min=&orig_term_min. , orig_term_max=&orig_term_max. ,wala_min=&wala_min. , wala_max=&wala_max. ,obsdt_bgn=&obsdt_bgn., obsdt_end=&obsdt_end. ,vintage_bgn=&vintage_begin. , vintage_end=&vintage_end. ,agency_typ=&agency_type. ,product_typ=&product_type. ,coll_typ=&coll_type. ,exclude_mon=&exclude_mon. ); /*read parameters initial values */ proc import datafile="&Parm_inputda_dir./&tunParmInFile." out=parmset DBMS=csv REPLACE; GETNAMES=YES; guessingrows=100; RUN; data parmset; set parmset; product="&product_name."; Agency=compress("&agency_type.","()'"); date_begin="&obsdt_bgn."d; date_end="&obsdt_end."d; run_date=Today(); run_by="&sysuserid."; format date_begin mmddyy10. date_end mmddyy10. run_date mmddyy10. ; run; /*performe nlin tuning */ %let year_end=%substr(&obsdt_end.,3,7); %nlin_ssn(&model_num.,testda_ft,&nlinout_db.,&nlin_sum., parmset, outest_param,"&nlin_graph_nm.","&output_dir.",&product_name.,&obsdt_bgn.,&obsdt_end. ); proc export data=nlin_parms dbms=csv outfile="&output_dir./&tunParmOutFile." replace; run; %mend; /*******************************************************************************************************************************************************************/ /*******************************************************************************************************************************************************************/ %macro ProdMBSpred(obsdate_b, obsdate_e, parm_dir, output_dir, inputdb,macroset_csvFile); proc import datafile="&parm_dir./&macroset_csvFile." out=macroset DBMS=csv REPLACE; GETNAMES=YES; RUN; data macroset1; set macroset; macro_ind=_n_; run; data _null_; set macroset1 end=last; if last then call symput('N', put (macro_ind,3.)); run; %do i=1 %to &N.; proc sql noprint; select strip(macroVar_name), macroVar_value into : this_macro_name , : this_macro_value from macroset1 where macro_ind=&i; quit; data _null_; call symput(strip("&this_macro_name."),strip("&this_macro_value.")); run; %end; %let year_end=%substr(&obsdate_e.,3,7); /*add some prefix to some macro variables */ %let TunParm_inFile=%cmpres(&TunParmInFile.); %let TunParm_outFile=%cmpres(&year_end._&product_nm._&TunParmoutFile.); %let nlinout_dtbase=%cmpres(&nlinout_db._&product_nm._&year_end.); %let nlinsum_dtbase=%cmpres(&nlin_sum._&product_nm._&year_end.); %let TunSum_outFile=%cmpres(&year_end._&product_nm._&tunSumOutFile.); %let Nlingraph_nm=%cmpres(&year_end._&product_nm._&nlin_graph_nm.); %put &product_nm.>>> &nlinsum_dtbase.>>>&tunParm_outFile. >>>> &nlinout_dtbase. >>>>> &TunSum_outFile. >>>> &Nlingraph_nm.; %mbspred(product_name=&product_nm., inputsasdb= &inputdb. ,month_reset_min=&month_reset_min. , month_reset_max=&month_reset_max. ,resetperi_min=&firstResetAge_min. , resetperi_max=&firstResetAge_max. ,orig_term_min=&orig_term_min. , orig_term_max=&orig_term_max. ,wala_min=&wala_min. , wala_max=&wala_max. ,obsdt_bgn=&obsdate_b., obsdt_end=&obsdate_e. ,vintage_begin=&vintage_begin. , vintage_end=&vintage_end. ,agency_type=&agency_type., product_type=&product_type., coll_type=&coll_type. ,exclude_mon=&exclude_month. ,Parm_inputda_dir=&parm_dir., tunParmInFile=&tunParm_InFile. ,nlinout_db=&nlinout_dtbase. ,output_dir=&output_dir. ,nlin_graph_nm=&nlingraph_nm., tunParmOutFile=&tunParm_OutFile., tunSumOutFile=&tunSum_OutFile. ,model_num=&ModelNum., nlin_sum=&nlinsum_dtbase. ); proc datasets lib=work; delete macroset macroset1; run; %mend; /*******************************************************************************************************************************************************************/ **** DRIVER PGM BELOW ****; /*********************************************************************************************************************************/ /******* *******/ /******* Description: This program use csv files from BRI then conduct optimization tuning and output the result to excel file *******/ /******* *******/ /******* *******/ /******* Change log: first version 4/18/2017 Huibin Cheng *******/ /******* *******/ /********* second version 06/11/2018 by Haisu Ma ****************/ /**/ ;;;; %let codedir=/lppra00a0133/data/dev/haisu/ginnie_plpm/run18; %let inputda_dir=/lppra00a0133/data/dev/haisu/ginnie_plpm; %let parm_dir=/lppra00a0133/data/dev/haisu/ginnie_plpm/run18; %let outputda_dir=/lppra00a0133/data/dev/haisu/ginnie_plpm/run18; libname sasdb "&inputda_dir."; libname outdb "&outputda_dir."; %include "&codedir./Macro_hm_bspline_GN30_server_poisson.sas"; %let date_begin=31JAN2012; %let date_end=31JAN2019; %let fix30_inputsasdb=sasdb.est_gnma30; %let fix30GNSetCsv=PLPM_GN30_Setting.csv; /*for model 1, no seasing, only tune refinance */ %macro full_run(n_runs,obs_begin,obs_end); %do run = 1 %to &n_runs.; %put running cycle &run.; %let obs_b=%sysfunc(intnx(month,"&obs_begin."D,%eval((&run.-1)*3),e),date9.); %let obs_e=%sysfunc(intnx(month,"&obs_end."D,%eval((&run.-1)*3),e),date9.); %put &obs_b. &obs_e.; %ProdMBSpred(obsdate_b=&obs_b., obsdate_e=&obs_e., parm_dir=&parm_dir.,output_dir=&outputda_dir., inputdb=&fix30_inputsasdb.,macroset_csvFile=&fix30GNSetCsv.); %end; %mend; %full_run(1,&date_begin.,&date_end.);     ... View more

My team uses SAS  9.4 (TS1M3) on Linux 2.6.32-754.el6.x86_64 (LIN X64) platform.  Proc Compare has been a part of our production data validation each month.  We use counts of differences as shown in Proc Compare's (proc compare base=work.prior compare=work.current listequalvar listvar novalues) output in the Ndif column:   Variables with Unequal Values Variable                 Type  Len  Ndif   MaxDif  MissDif acquired_dt              NUM     8     3     4750        1 bkr_code                 CHAR    2  7495              7364 bkr_status               CHAR    1  7576              7318 orig_dt                  NUM     8     2      911        0 clnt_inv                 CHAR    6 65810                 0 corpadv_r                NUM     8 36885   117431        0 corpadv_r3               NUM     8197427  2151214        0 corr_data                CHAR    6 63001             63001 distrib_ty               CHAR    1  1272                 0 esc_doc_ty               CHAR    2    37                 0 escr_bal                 NUM     86.37E6   229448        0 ...   Note however that there is no space or delimiter between "Len" and "Ndif".  My team's SAS code to parse this produced incorrect values today since it couldn't distinguish between the "8" for Len, and the remaining count (Ndif).  Yes, there are several ways to work around this, but the fact is, it is a bug.  SAS should fix it.  They should have fixed it long ago. ... View more

I'm running on Linux 64, 9.4 (TS1M3).  I can produce an SGPLOT vbar chart and do a Proc Print of a Proc Means... but the output for the SGPLOt is an hour older than that of the Print.  See attachment.  Can't find anything on this in the SAS documentation.  I really don't understand ODS and am just copying things that I've seen others do.         ODS HTML PATH = "&outdir/" (url=none); **ods graphics /imagemap=on; ods html file="&report_file..html"; title "BRI Jobs Linux Space - &Server, %sysfunc(today(),weekdate3. -r)%str(,) %sysfunc(today(),worddate15. -l) %sysfunc(time(),tod5. -l)"; footnote h=3pt "%sysfunc(getoption(sysin)), &sysmacroname"; proc means data=work.bri_space_report nway noprint; class user; var SizeMB; output out=work.bri_space_sums sum(SizeMB)=; run; *****************************************************; * Generate the vertical bar plot by user; *****************************************************; proc sgplot data=work.bri_space_sums; vbar user / response=sizeMB; run; proc sort data=work.bri_space_sums; by descending SizeMB; run; proc print data=work.bri_space_sums noobs; format _numeric_ comma10.; var user _freq_ sizemb; run;   ... View more

Here's the background.  I'm on Linux 9.4 (TS1M3) and I want to run a set of my users' typical queries via SAS/Access for Oracle from a cron job, then store the results (in perm SAS file) hourly to monitor performance over time.  The first query I have started with has a couple variables that may change next year, so I decided to put it into a macro with parameters for those vars.  The macro(s) will be called from within the Proc SQL block.   But I'm running into a lot of unexpected issues.  Apparently SAS/Access for Oracle (despite SAS's rich annual fees) doesn't recognize Oracle SQL very well.  For example,   select clean_book_ind, book_type,     case when gov_ind = 'F' and risk_type_cd not in ('FMG','FNG','GNF','GNV','GNR') then 'FHAP'     else 'OTHER'     end private_ind, sum(balance) upb ... (etc)   had to be changed to include "as" for the tags:   select clean book_ind, book_type,     case when gov_ind = 'F' and risk_type_cd not in ('FMG','FNG','GNF','GNV','GNR') then 'FHAP'     else 'OTHER'     end AS private_ind, sum(balance) AS upb ... (etc)   so for a lengthy query there is a lot of manual re-work. Then I couldn't pass in a single-ticks date: where asset_date = %bquote('&asset_date')... (it was '31jan2018) I had to use %unquote(%bquote('&asset_date')).   Next, having cleaned up those issues, I got ERROR: ORACLE prepare error: ORA-24333: zero iteration count.   Googled it, got to Usage Note 15432: "ORA-24333: zero iteration count" error, tried that kind of syntax, (proc sql;    connect to oracle (user=scott pw=tiger);    execute (      create table newemp as select * from emp    ) by oracle;    disconnect from oracle;  quit;) (unfortunately no "create table work.blahblah as select * from connection to oracle"  lines)   and am now getting "ERROR: ORACLE execute error: ORA-00900: invalid SQL statement" but then I copied the SQL statement from my SAS log, pasted it into my Windows SQL Client (Squirrel) and it ran just fine.   Yep, I am frustrated with this product - it's not living up to the reputation it *should* have for the money we are paying - and I don't know what to do next.   ... View more