About PhilC

PhilC · ‎02-11-2021

Thanks for your time @Ksharp. Hash tables are probably the best way to do this given all of my limitations, but, more than this, the coders I work here with are not ready for hash tables, to be honest. I want to avoid it. But I did. My original problem was a time period overlap problem, and I had experience working in other platforms with graph theory and network propagation so I found my self asking this question. Work you did with @yabwon in September 2019 helped me in the end. thanks again.

PhilC · ‎02-10-2021

Again, not simple, but there is the parsing of logs. I have seen others do this. The process flow name is put in the log when the variable &_CLIENTPROCESSFLOWNAME is assigned. In theory you can link that to datasets and time stamps, but not exact time stamps, unfortunately.

PhilC · ‎02-10-2021

Thanks for the inspiration! This helps me too. But I'm pretty sure, so math works in every case, that we should sort by your marker variable. Incase we have Start and End dates that coincide and are given in HAVE in the wrong order, we do want the accumulator to be added to before being subtracted. proc sort data = have2; by ID Date /**/ descending marker /**/; run;

PhilC · ‎02-10-2021

I have no simple answers. If I wanted to do this I would have, planning ahead, created LIBNAMEs for each process flow. And then used PROC SQL and DICTIONARY.TABLES to generate the table, you speak of. Or named the datasets prefixed with the macro variable &_CLIENTPROCESSFLOWNAME. Which would have required me to keep process flow names simple. I hope you find a simpler idea. Is it that you want to know If there was an error in the process flow that generated the "AUTOLOAD" dataset and the timing of that error, which can be compared with the dataset's modified date?

PhilC · ‎02-09-2021

I think I need more information. I'm speculating that you have opened a Enterprise Guide file (EPG) and that you see "tables" (ie. datasets) in a process flow in that file, but you receive an error when you attempt to view the dataset via the dataset icon. Correct me, if I misunderstand you.

PhilC · ‎02-09-2021

I have data serialized by a variable N, and I have a dataset that is virtually an incidence matrix. How might I find the eleven (11) groupings of records in HAVE that are connected to common undirected graphs defined by the incidence_matrix dataset? Caveats: I only have SAS STAT. Using any other SAS package is not an option, but please include your idea anyway. Do not assume big data. Expect less than 10k rows for HAVE in real life, and a matrix about as sparse. data have ; do N=1 to 15; output; end; stop; data incidence_matrix(index=(N)); input N N2; datalines; 1 2 2 10 5 6 5 7 ; run; want: a column that groups the records. (Answering @Reeza) User should be able to perform sorts and use by statements with the new column. An example dataset: N X 1 20 2 20 3 3 4 4 5 30 6 30 7 30 8 8 9 9 10 20 : : 15 15

PhilC · ‎02-08-2021

Now I see. If you care to understand more about this subject in SAS in the future, a good term to remember is SAS "Program data vector" or PDV. Which is what I would have spoken to if I saw what you were dealing with in time. There was a similar question asked about ordering of variables in the last two weeks.

PhilC · ‎02-08-2021

OK, question then, if one might imagine the variables B1 through B5. Would one want, then, variables B6 through B30? That is are you wanting to keep the same prefix? in your case "V" and in this case "B".

PhilC · ‎02-08-2021

What's your previous experience with programming? I am curious. you can do this the "wide" way or the "narrow" way. the wide: data want_wide; year1 =1988; year2 =1999; year3 =2000; run; the narrow: data want_narrow; year =1988; output; year =1999; output; year =2000; output; run;

PhilC · ‎02-08-2021

I would need more information, as well. but to start to answer your questions: 1) This is not an Enterprise Guide question, it is a SAS programming question. 2) One can add column using this syntax. data want; set have; length v8-v30 8; run; that's going to add 23 numeric columns to your data. 3) This code will access the metadata of the dataset "HAVE" and create a table, one row per column. proc sql; select * from dictionary.Columns where libname="WORK" AND memname="HAVE" ; quit; using this type of query one can "count" how many columns are in a dataset.

PhilC · ‎02-05-2021

Hi all, related question for my own knowledge. Is there a downside in using DoSubl() for the OP's code? My naive opinion is that it would fix the timing.

PhilC · ‎02-05-2021

This is could be handled using Excel's Find and Replace. ?

PhilC · ‎02-05-2021

That's heavily dependent on your operating system, and if you have the correct privileges granted by your organization's IT.

PhilC · ‎02-05-2021

This isn't a join, as far as I understand SQL notation and syntax. This can be achieved in SQL using a UNION but it requires another pass to get the order correct. But SAS can interleave the records; I was taught the term collate, I digress. The datasets A and B I presume are properly sorted. data want; set a b; by id date; run;

PhilC · ‎02-05-2021

This is where some SAS is interjected into SQL syntax for PROC SQL, namely the SAS dataset options. Try this replace this /*...*/ from PSAEU11.db_dossier sousc (where sousc.cd_dossier = 'SOUSC' and sousc.lp_etat_doss not in ('ANNUL','A30','IMPAY') and sousc.no_police = 'I99300001') left join /*...*/ with this /*...*/ from PSAEU11.db_dossier (where=(cd_dossier = 'SOUSC' and lp_etat_doss not in ('ANNUL','A30','IMPAY') and no_police = 'I99300001')) as sousc left join /*...*/ This little example works as well proc sql; select cars.Make, cars.model from SASHELP.CARS(where=(Make="Audi")) as cars; quit;

Online Status	Offline
Date Last Visited	‎04-12-2022 11:06 PM

Re: Creating a categorical variable conditioned on a series of if-then...

Re: Creating a categorical variable conditioned on a series of if-then...

Re: Creating a categorical variable conditioned on a series of if-then...

Re: Add Excel File Name as Variable When Importing Multiple Excel File...

Re: How do I format a field using two different formats in proc report

Re: Allow PROC SORT to output multiple datasets

Re: Allow PROC SORT to output multiple datasets

Re: Looking for Like Values in Array

Re: Find Outliers and Matches in dataset

Re: Need help with code, calculating proportion of days covered by med...

Re: How do I format a field using two different formats in proc report

SAS EG 8 scroll bars

Support WITH clauses in SQL

Re: Support WITH clauses in SQL

EG 8 - Bring back the RUN command on the context menu

Re: After upgrading to 8.2, EG could not "find" local host

Re: Option to save the LOG Summary to a dataset

Re: ODS EXCEL: option for the inserting of line feeds when wrapping li...

Re: XML Mapper exporting multiple tables

Re: Looking for Like Values in Array

Re: Clustering with an Incidence Matrix

Re: Automatically get table names executed in a flow in SAS Enterprise...

Re: Combine overlapping dates while keeping non-continuous intervals s...

Re: Automatically get table names executed in a flow in SAS Enterprise...

Re: Automatically get table names executed in a flow in SAS Enterprise...

Clustering with an Incidence Matrix

Re: Adding columns based on condition

Re: Adding columns based on condition

Re: dataset with variable that is a list of strings

Re: Adding columns based on condition

Re: Data step question

Re: Any 'cell' in database NOT CONTAINING a %... change to missing

Re: Schedule a script with program

Re: join table by id

Re: proc sql: the where clause