About ibsulkim

ibsulkim · ‎03-24-2021

Thank you very much Kathleen for prompt help and for sharing the links - they are all very helpful. Proc mixed seems a solution in my case, and thanks for the link that directly compares between proc panel and proc mixed. All the best.

ibsulkim · ‎03-24-2021

Hello, I have a panel dataset with a number of firms over different dates. Assume the left hand side variable in the panel regression is called y. I am trying to run a panel regression using the inverse of the std.dev(y) as regression weight (and include time and firm fixed effects). Could anyone help how I can implement this in SAS? I have been using proc panel to include the time and firm fixed effects as below, and I was hoping to add "weight" option in the code (similar to what we do in proc reg), but it seems there is no such option available for proc panel. proc panel data=data PRINTFIXED; id firm_id date; model y = x1 x2 x3 / fixtwo ; output out=est residual=residual predicted=predicted; run; Any help would be greatly appreciated.

ibsulkim · ‎08-10-2018

Hello, How can I have my own weight matrix for moment conditions under proc model gmm? There is "Weight" statement under proc model, however, it gives weight only to each observation. I would like to have weight for each moment condition in GMM but I couldn't find any option to do it under proc model gmm. Thank you.

ibsulkim · ‎03-17-2017

Hello Everyone, In one-way firm fixed effect panel regressioin, how can I get heteroskadasticity-robust standard errors clustered by quarter? I would like the standard errors to be clustered by quarter, but based on the SAS support (link) it seems I can only get heteroskadasticity-robust standard errors clustered by firm, not by quarter. The code I have in mind is proc panel data=have plot=all PRINTFIXED; id firm quarter; model Y = x1 x2 / fixone noint HCCME=3 cluster; output out=est_panel residual=residual predicted=predicted; run; and when I run this code, I believe the standard error I get is clustered by firm, and I canot find a way to cluster it by quarter. In Stata, I can specify by which variable I would like the standard errors to be clustered but I can't find the similar option in SAS. Could anyone help? All the best, Minsoo

ibsulkim · ‎01-06-2013

Thanks Arthur, it just works perfect!

ibsulkim · ‎01-06-2013

Hi, I am trying to work on a large dataset in total of approximately 40TB. I can access to the datasets by connecting to a server that has all the datasets saved as sas dataset format. Here I am trying to find an efficient coding that reduces running time as much as possible. The only output I need is the group difference for a couple of variables such as 'count' and the group mean such as 'meanspread' that I compute using proc means at the end of the attached code. In order to reduce the time, I tried to use 'rsubmit' and 'SAS Data Views'. However, it still takes way too long and any comment or feedback that helps to reduce the running time is truly appreciated at this moment. FYI, the name of the dataset in the remote library 'taq' is composed of two parts: cq_(or ct_) + date for the dataset, e.g., cq_19930102, cq_19930103,... : libname mydir 'C:\sasdata\taq'; %let wrds=wrds.wharton.upenn.edu 4016; options comamid=TCP remote=WRDS; signon username=_prompt_; rsubmit; libname taq '/wrds/taq/sasdata'; data quotes/view=quotes; set taq.cq_: ; run; data trades/view=trades; set taq.ct_: ; run; data tradesquotes/view=tradesquotes; set trades quotes ; by symbol date time type; run; data tr_qt/view= tr_qt; set tradesquotes; lag_date=lag(date); if date^=lag_date or symbol^=lag_symbol then do; retain prior_qtime_adjusted; if type='T' then output; keep symbol date time price size prior_bid prior_ofr; run; data tr_qt2/view= tr_qt2; set tr_qt; lagprice1=lag(price); mid=(prior_ofr+prior_bid)/2; if price>mid then order=1; else if price<mid then order=-1; run; proc means data=tr_qt2 noprint; var size dollars; class month order; output out=output_raw n=count sum(size dollars)=sumsize sumdollars mean(spread)=meanspread; run; data outputcut; set output_raw; where _type_=3; run; proc transpose data=outputcut out=output_num; by month; var count; run; proc transpose data=outputcut out=output_sh; by month; var sumsize; run; data tmp_OIBNUM(keep=month OIBNUM); set output_num; OIBNUM=COL3-COL1; run; data tmp_OIBSH(keep=month OIBSH); set output_sh; OIBSH=COL3-COL1; run; data tmp_LIQ (keep=month QSPR NUMTRANS VOL); set output_raw; where _type_=2; QSPR=meanspread; NUMTRANS=count; VOL=sumdollars; run; data finalout; merge tmp_OIBNUM tmp_OIBSH tmp_LIQ; by month; run; proc download data=finalout out=mydir.final; run; proc print data=finalout (obs=100); run; endrsubmit;

ibsulkim · ‎01-06-2013

Hi, I am using SAS 9.3 and would like to use proc transpose for the following dataset named as outputcut: [outputcut] month order count 200701 -1 209 200701 0 100 200701 1 89 200702 -1 324 200702 1 291 When I run the following code: proc transpose data=outputcut out=output_num; by month; var count; run; it gives me the output as follow: month _NAME_ COL1 COL2 COL3 200701 count 209 100 89 200701 count 324 291 . However, I would like to get my output as follow: month _NAME_ COL1 COL2 COL3 200701 count 209 100 89 200701 count 324 . 291 so that COL2 corresponds to the value 0 of the variable 'order', and COL3 to the value of 1 for the variable 'order'. Any comment is appreciated! Minsoo

ibsulkim · ‎08-02-2011

Great! You make it happen! Thank you very much for your solutions, Tom and Art! Minsoo

ibsulkim · ‎08-02-2011

Hi Tom, Yes, your comment seems to be the reason why it shows me the warning message. It also makes sense that SAS doesn't make warning message when the case sentitive variable names are exactly the same, however, it gives an warning message when the variable has the same spell but different case for the variable names. Is there any way I can go around this warning message except chaning the variable names into another one? Thanks, Minsoo

ibsulkim · ‎08-02-2011

I need want file to have CompnayName, TalbeName, Date, Year on the first 4 columns and then have various variables such as Roda_and_equipemnt, etc. on the following columns as below: CompanyName TableName Date Year Road_and_equipment Invest__in_affil__cos ... ..... .... The previous code works fine except for the variables shown in the warning messages. Thanks, Minsoo

ibsulkim · ‎08-02-2011

Hi Art, The have dataset I have posted is the first few lines of the dataset, and yes, it works fine with it, however it keeps making the warning message if I run it with the entire dataset. The following link shows you the dataset. And here is the code that I have used for this dataset: PROC IMPORT OUT= WORK.have DATAFILE= "C:\Users\.......\1916_b_Macro.xls" DBMS=EXCEL REPLACE; SHEET="Sheet1$"; GETNAMES=NO; MIXED=YES; SCANTEXT=YES; USEDATE=YES; SCANTIME=YES; RUN; data have1; set have; array vars(100) $200.; array vars_in(100) $200. f1-f100; retain vars; if compress(f1,,'kfd') eq "CompanyName" then do; call missing(of vars(*)); do i=1 to 100; if compress(vars_in(i),,'kfd') ne "" then vars(i)=vars_in(i); end; end; else if compress(f1,,'kf') ne "" then do; CompanyName=vars_in(1); Date=vars_in(2); TableName=vars_in(3); Year=vars_in(4); do i=5 to 100; if compress(vars_in(i),,'kd') ne "" then do; var=vars(i); data=vars_in(i); output; end; end; end; run; proc sort data=have1; by CompanyName TableName Date Year; run; proc transpose data=have1 out=want (drop=_NAME_); by CompanyName TableName Date Year; var data; id var; run; Thanks, Minsoo

ibsulkim · ‎08-02-2011

Hi, I get warning messages in log when I run the following codes with the data: dataset- have, can be simplified as below: CompanyName Date TableName Year var data ALABAMA & VICKSBURG RY 6026 Assets- 1914 Road and equipment 5367994 ALABAMA & VICKSBURG RY 6026 Assets- 1914 Invest, in affil. cos 34700 ALABAMA & VICKSBURG RY 6026 Assets- 1914 Misc. physical property 1580 ALABAMA & VICKSBURG RY 6026 Assets- 1914 Other Investments 50 ALABAMA & VICKSBURG RY 6026 Assets- 1914 Cash 754253 ALABAMA & VICKSBURG RY 6026 Assets- 1914 Marketable securities 4247 proc sort data=have; by CompanyName TableName Date Year; run; proc transpose data=have out=want (drop=_NAME_); by CompanyName TableName Date Year; var data; id var; run; The warning message is as follow: WARNING: Variable Cash already exists on file WORK.WANT. WARNING: Variable Total_Income already exists on file WORK.WANT WARNING: Variable Net_earnings already exists on file WORK.WANT WARNING: Variable Other_income already exists on file WORK.WANT. . . . . . . . . From one of SAS support document, I have read that there should be no duplicate values in id variable within a by group. and I could take care of the the following error message: ERROR: The ID value "PassengerEarnings" occurs twice in the same BY group. However, I couldn't figure it out why SAS shows me the warning message above. Any comment or suggestion will be greatly appreciated. Thanks, Minsoo

ibsulkim · ‎07-19-2011

Hi Art, Great! It works just fine! Your help is tremendously valuable for me. Thanks, Minsoo

ibsulkim · ‎07-17-2011

Art, I didn't fully understand your code at the moment I replied in that your code automatically identifies the identical variable in columns and put body of contents in the right order. Your code brings me up some possibility to reduce tremendous amount of time if it can be applied to my data after some of modification. Looking forward to seeing this happens! One remaining issue is that my data actually have character variables in the table headers unlikely V1-V14, and I can't manually count how many variables I have in total because it has too many tables. Could you make any follow-up suggestion or comments on these issues? Any of your comments will be greatly appreciated. Thank you. Minsoo

ibsulkim · ‎07-15-2011

Hi, I believe there must be a simple way to copy value in the next row(observation) within the same column. Data set can be simplifies as below: [original data set] V1 V2 V3 V4 V5 aa bb cc dd ee mk jk as rd as aa bb cc dd ee re te sf er qw [goal data set] V1 V2 V3 V4 V5 mk bb as dd ee mk jk as rd as re bb sf dd ee re te sf er qw I would like to change the value 'aa' into the value in the next row and change the value 'cc' into the value in the next row. Any comment or suggestion will be greatly appreciated. Thanks, Minsoo

Online Status	Offline
Date Last Visited	‎09-21-2021 03:22 AM

Re: Proc Panel with regression weight

Proc Panel with regression weight

proc model gmm weight

Clustered standard errors in proc panel

Re: proc transpose with different number of by values

reduce running time that works for a large dataset saved in a remote l...

proc transpose with different number of by values

Re: warning in proc transpose

Re: warning in proc transpose

Re: warning in proc transpose

Re: Proc Panel with regression weight

Re: Proc Panel with regression weight

Re: Proc Panel with regression weight

Re: Proc Panel with regression weight

how to merge table by table

Re: Proc Panel with regression weight

Proc Panel with regression weight

proc model gmm weight

Clustered standard errors in proc panel

Re: proc transpose with different number of by values

reduce running time that works for a large dataset saved in a remote l...

proc transpose with different number of by values

Re: warning in proc transpose

Re: warning in proc transpose

Re: warning in proc transpose

Re: warning in proc transpose

warning in proc transpose

Re: how to merge table by table

Re: how to merge table by table

copying value from the next observation