About cstarkey

cstarkey · ‎03-25-2021

Brilliant, thank you very much! I've mainly dealt with Excel files so Proc Import has always been the go to. The infile method you showed me not only fixed my issues with the dates but also the transpose is no longer even needed. I can't thank you enough!

cstarkey · ‎03-25-2021

Month 1 = Jan, month 2 = Feb, etc...

cstarkey · ‎03-24-2021

Hello. I posted a question about transposing this csv dataset and now I'm stuck at trying to get the date to format correctly. As before, the dataset is like: Country Counter_Country Type 2001M1 2001M2 2001M3 France US Import 10 11 12 France US Export 4 5 6 France UK Import 1 2 3 France UK Export 2 3 4 US FR IM 3 4 5 US FR EX 5 6 7 US UK IM 4 5 6 US UK EX 5 6 7 My code to import and transpose is: PROC IMPORT DATAFILE= "C:\file\dot.csv" OUT= dot_Data DBMS=csv; GETNAMES=yes; proc sort data=dot_data; by Country Counter_Country type; RUN; proc transpose data=dot_Data out=dot_long; by Country Counter_Country type; *by category notsorted; run; This takes the dates (1985M1, 1985M2, etc...) from being variable names and makes a column with the default _NAMES_. The data is one column: Country Counter_Country Type _NAME_ COL1 France US Import 2001M1 10 France US Import 2001M2 11 France US Import 2001M3 12 France US Export 2001M1 4 France US Export 2001M2 5 France US Export 2001M3 6 etc... It makes the _NAME_ col format character $8. I need to make it a date format. Amongst other attempts, I've tried: data dot_temp (rename=(COL1 = Value)); set dot_long; dte= input(_NAME_,anydtdte32.); format dte YYMON7.; run; but it only create a dte column with a period. Is there any way to import the data as a date format or to change it to a date format? Any help would be appreciated. Thank you.

cstarkey · ‎03-19-2021

Thanks for your help, I really appreciate it. My problems transposing it were making me overcomplicate it. That by statement is exactly what I need. Thanks to you and Reeza for the help!

cstarkey · ‎03-19-2021

Thanks for the replies. An example would be something like: Country Counter Country Type Date 2001-01 2001-02 2001-03 France US Import 10 11 12 France US Export 4 5 6 France UK Import 1 2 3 France UK Export 2 3 4 US FR IM 3 4 5 US FR EX 5 6 7 US UK IM 4 5 6 US UK EX 5 6 7 UK FR IM UK FR EX UK US IM UK US EX This goes on for over 100 countries with numerous variables, so renaming the labels to something like US_UK_IM isn't ideal. The idea is that I can sort by country, counter country, or type while also restricting the data to a country (such as the US for the financial crisis, China for the 2015 market event, etc) and mix those events together for my tests. Ideally I could transpose the data set to make it long with the labels being nested so I could restrict the data to something like both the US being the country and the counter country. Please let me know if this explanation is helpful or not. Thank you

cstarkey · ‎03-18-2021

Hello. I have a time series data set for imports and exports from IMF. It has column one as the main country, column two as the counter country, column three as the explanation (import or export), and the rest of the columns are values, with each column representing a month. I'm trying to transpose it but it's difficult with the 3 explanation variables (country, counter, and type). It seems that having a multi-dimensional label would be great but I haven't found anything making me think it's possible. Of course I could combine the columns (ie country-counter-type) but this makes it difficult when I need to sort by crisis country. Any suggestions would be appreciated. If Python or another language offers this, that information would be helpful also. Thank you.

cstarkey · ‎01-02-2020

Thanks for the response. Here is the code for varratio, which doesn't run: proc psmatch data=event region=treated; class HB st; psdata ps = ps_min treatvar=HB(treated='1'); match distance=ps method=varratio(kmin=1 kmax=2) exact=(st) caliper=.; assess ps var=(A B C D E F) / plots(orient=vertical); output out(obs=match)=match_varratio_event matchid=_MatchID; run; And the code for greedy, which runs fine: proc psmatch data=event region=treated; class HB st; psdata ps = ps_min treatvar=HB(treated='1'); match distance=ps method=greedy(k=5) exact=(st) caliper=0.25; assess ps var=(A B C D E F) / plots(orient=vertical); output out(obs=match)=match_greedy_event matchid=_MatchID; run; Thank you.

cstarkey · ‎01-01-2020

Thanks for the suggestion. I'll try reducing the amount of rows, though that doesn't really help the analysis long term. There's about 242k rows now. Still, it will give me a way to try and figure out where the problem is. Greedy match doesn't have a problem with the size so I'm at a loss why any other method stops so abruptly. The thing that really confuses me is why the fullstimer stats don't reflect what my system has, in terms of memory.

cstarkey · ‎12-31-2019

Hello. Thanks in advance to anyone who can help me. I'm trying to run a psmatch procedure with pscores already in the dataset. I can run a greedy match fine but if I try to do a varratio or mah distance I get the insufficient memory error. I've tried different values of memsize, finally setting it to 0, but still get the error. I have 32G of ram with large amounts of hd space for virtual memory. The last run of varratio, kmin=1 kmax=2, returned the error with: ERROR: The SAS System stopped processing this step because of insufficient memory. WARNING: The data set WORK.MATCH_GREEDY_EVENT may be incomplete. When this step was stopped there were 0 observations and 159 variables. NOTE: PROCEDURE PSMATCH used (Total process time): real time 5:16.59 user cpu time 3:08.21 system cpu time 2:08.23 memory 7390631.12k OS Memory 7410196.00k Timestamp 12/31/2019 09:29:59 PM Step Count 2 Switch Count 3 This is despite: Group=MEMORY SORTSIZE=1073741824 Specifies the amount of memory that is available to the SORT procedure. SUMSIZE=0 Specifies a limit on the amount of memory that is available for data summarization procedures when class variables are active. MAXMEMQUERY=0 Specifies the maximum amount of memory that is allocated for procedures. MEMBLKSZ=16777216 Specifies the memory block size for Windows memory-based libraries. MEMMAXSZ=2147483648 Specifies the maximum amount of memory to allocate for using memory-based libraries. LOADMEMSIZE=0 Specifies a suggested amount of memory that is needed for executable programs loaded by SAS. MEMSIZE=22438694400 Specifies the limit on the amount of virtual memory that can be used during a SAS session. REALMEMSIZE=0 Specifies the amount of real memory SAS can expect to allocate. NOTE: PROCEDURE OPTIONS used (Total process time): real time 0.00 seconds user cpu time 0.00 seconds system cpu time 0.00 seconds memory 198.21k OS Memory 16872.00k Timestamp 12/31/2019 09:38:27 PM Step Count 1 Switch Count 0 3 4 %put Free RAM = %sysfunc(putn(%sysfunc(getoption(xmrlmem))/1024**2,comma10.))MB; Free RAM = 17,106MB Again, this is with memsize set to 0. I've even tried upping the amount of virtual memory so I had about 60G total but set it back to let windows manage it. If anyone can help with a suggestion, I'd really appreciate it. Also, if someone can explain with fullstimer stops at 7G when there is still a lot of slack that would really help me understand how SAS and windows are managing my memory. Thank you.

Online Status	Offline
Date Last Visited	‎05-07-2021 06:49 PM

Re: import csv YYMM date

Re: import csv YYMM date

import csv YYMM date

Re: multi-label for time series

Re: multi-label for time series

multi-label for time series

Re: Insufficient Memory Error

Re: Insufficient Memory Error

Insufficient Memory Error

Re: import csv YYMM date

Re: multi-label for time series

Re: multi-label for time series

Re: multi-label for time series

Re: import csv YYMM date

Re: import csv YYMM date

import csv YYMM date

Re: multi-label for time series

Re: multi-label for time series

multi-label for time series

Re: Insufficient Memory Error

Re: Insufficient Memory Error

Insufficient Memory Error