[pre] ods pdf file='c:\test.pdf' ; proc report data=sashelp.class nowd style(report)={rules=none frame=void outputwidth=50%} ; column name sex age weight height; define name /order noprint; define sex /order noprint; define age /order noprint; define weight /order noprint; define height/order noprint; compute before height; line @8 name $10. @20 sex $1. @25 age 2.; line @10 weight 5.2 @20 height 5.2; endcomp; run; ods pdf close; [/pre] Ksharp ... View more

>whenever I use informat for character variable in the input statement, SAS will treat it as formatted input, unless I use colon. It makes a real difference when my values have different length. Yes. you are right. But the length of variable is the length you defined in format, that is not different length when you use colon in your input statement.There is an important thing you need to remember(i.e. when you use length statement and informat statement before input statement, input method is identical with colon input method just as Cynthia mentioned),and once the character varible enters the PDV ,its length will not allow to change afer data step. Now let's take a look at your example. In your code ,'hours'is list input, 'Firstname' is list input(which has eight length sas default),'Lastname' is formatted input ( which will ignore the delimiter such as blank, and input until the fifth character, So you will get 'Lee 1' not 'Lee', you should add colon before $5. such as : $5.), The colon in ' : $5. ' will stop read the data when encounter delimiter ( blank and so on),but the length of variable is still 5. Hope this will help you a little bit. Cynthia gives some value reference about it. Ksharp Message was edited by: Ksharp ... View more

Sorry. I have not such experience.Maybe Google would help you a little bit. Or just write some UNIX shell script. ... View more

Hi.Peter. >Have you ever created so many tables in one data step? I have not created so many datasets before.But I used the method you mentions to label the variables in lots of datasets before,finally the macro variable &outputs overflows the max length of itself under UNIX environment.So I have to sparse the &outputs as I mentioned. And your gived example is very interesting and useful for me. Admittedly this way is very to waste OS source,Also hope someone can give more optimal alternative. About @data _null_;?? It is long time no see him/her and Mr.Patrick. Ksharp ... View more

Hi. I try to use proc report's line statement to get what you want. But I do not know it is whether you exactly needed. proc report's line statement is very powerful and useful. [pre] ods pdf file='c:\test.pdf' style=sasweb; proc report data=sashelp.class nowd ; column name sex age weight height; define name /order noprint; define sex /order noprint; define age /order noprint; define weight /order noprint; define height/order noprint; compute before height; line @8 name $10. @20 sex $1. @25 age 2.; line @10 weight 5.2 @20 height 5.2; endcomp; run; ods pdf close; [/pre] Ksharp ... View more

Sure.Of course. E-mail: xiakeshan@yahoo.com.cn But I am in china now. And maybe you need some more time to consider for your whole data. I can explain my code if you has some uncertainty. Try it at your best. So I think you can do it by yourself. Good Luck. [pre] 19/03/2004 100174701 BE0003565737 44,8 buy 2300 6/03/2007 100174701 BE0003565737 89,67 sell 2300 24/03/2004 100174701 BE0003565737 44,59 buy 12050 6/03/2007 100174701 BE0003565737 89,67 sell 37700 23/06/2005 100174701 BE0003565737 66,65 buy 3900 6/03/2007 100174701 BE0003565737 89,67 sell 33800 [/pre] 33800=37700-3900 not 37700-12050 Is it rigth? Message was edited by: Ksharp ... View more

>if character values are shorter than 8 bytes and do not have embedded blanks, the modifiers are not necessary. That is only suitable for list input .that is mean list input ' input a $ ' is the same as modified input ' input a : $8.' .Once you use formatted then maybe you would use colon as Art mentioned ' input a : $7.'. ... View more

Hi. I have another choice. [pre] data one; input names $; datalines; XDANO XXQRX XMKR XXXDMZ CANE ; run; data one;; set one; if upcase(names) eq : 'X' then names=substr(names,verify(names,'xX')); put names=; run; *or just more explicit and simple; data one;; set one; names=substr(names,verify(names,'xX')); put names=; run; [/pre] Ksharp Message was edited by: Ksharp ... View more

Yes.Peter. You have mentioned a problem in your code. Macro variable has limited length (32767 or something else),That is to say when &outputs can not hold all the value from PROC SQL, will trigger an error. Ksharp. ... View more

Sorry.The code above has some problem. The following code is corrected version. P.S. The code is recommended for the sake of your partial data. [pre] data i; input tdate : DDMMYY10. fund : $10. ISIN : $20. Price : commaX8. transmod $ numberstocks; format tdate DDMMYY10. price commaX8.2; datalines; 19/02/2004 100174701 BE0003565737 47 buy 3000 19/03/2004 100174701 BE0003565737 44,8 buy 3800 24/03/2004 100174701 BE0003565737 44,59 buy 12050 23/06/2005 100174701 BE0003565737 66,65 buy 3900 11/11/2005 100174701 BE0003565737 73,15 buy 3000 14/11/2005 100174701 BE0003565737 73,2 buy 3800 17/05/2006 100174701 BE0003565737 84,3 buy 250 12/06/2006 100174701 BE0003565737 78,05 buy 800 11/08/2003 100174701 BE0003565737 87,85 sell 40000 11/08/2006 100174701 BE0003565737 87,85 sell 3800 21/09/2006 100174701 BE0003565737 84,75 buy 530 23/11/2006 100174701 BE0003565737 85,3 buy 30 4/12/2006 100174701 BE0003565737 85,95 buy 20 23/01/2007 100174701 BE0003565737 97,65 buy 260 13/02/2007 100174701 BE0003565737 97,8 sell 400 23/02/2007 100174701 BE0003565737 96,2 buy 150 6/03/2007 100174701 BE0003565737 89,67 sell 300 6/03/2007 100174701 BE0003565737 89,67 sell 40000 ; run; proc sort data=i ; by tdate ; run; data i; set i; if _n_=1 then delete; run; data buy sell(rename=(tdate=_tdate fund=_fund isin=_isin price=_price transmod=_transmod numberstocks=_numberstocks)); set i; if transmod eq 'buy' then output buy; else if transmod eq 'sell' then output sell; run; data want; b=1;s=1; set buy nobs=b_obs; set sell nobs=s_obs; sum+numberstocks; _sum+_numberstocks;b_flag=0;s_flag=1; output; do while(b lt b_obs or s lt s_obs); do while(sum lt _sum); b+1; set buy point=b; sum+numberstocks;b_flag=0; s_flag=1; output; if b ge b_obs and s ge s_obs then stop; end; do while(sum gt _sum); s+1; set sell point=s; _sum+_numberstocks;b_flag=1; s_flag=0; output; if b ge b_obs and s ge s_obs then stop; end; end; run; data op; set want; dif_sum=_sum - sum; run; data opp; set op; lag_dif_sum=abs(lag(dif_sum)); drop sum _sum dif_sum ; run; data temp; set opp; if _n_ eq 1 then _numberstocks = numberstocks; else do; if b_flag =0 and s_flag=1 then _numberstocks=lag_dif_sum; else if b_flag=1 and s_flag=0 then numberstocks=lag_dif_sum; end; run; data result; set temp; if b_flag=1 and s_flag=0 and _numberstocks gt lag_dif_sum then _numberstocks =numberstocks; drop b_flag s_flag lag_dif_sum; run; proc print noobs;run; [/pre] Ksharp Message was edited by: Ksharp Message was edited by: Ksharp ... View more

Yes.I will give you some help. I refined my code again. The following code is outcome I have worked a whole day ,I think I am very tired. I have get the result looks like yours. But due to your partial data (not complete) and complicated logic, so this code is only reference. If you have some other problem ,can post it. Good Luck. [pre] data i; input tdate : DDMMYY10. fund : $10. ISIN : $20. Price : commaX8. transmod $ numberstocks; format tdate DDMMYY10. price commaX8.2; datalines; 19/02/2004 100174701 BE0003565737 47 buy 3000 19/03/2004 100174701 BE0003565737 44,8 buy 3800 24/03/2004 100174701 BE0003565737 44,59 buy 12050 23/06/2005 100174701 BE0003565737 66,65 buy 3900 11/11/2005 100174701 BE0003565737 73,15 buy 3000 14/11/2005 100174701 BE0003565737 73,2 buy 3800 17/05/2006 100174701 BE0003565737 84,3 buy 250 12/06/2006 100174701 BE0003565737 78,05 buy 800 11/08/2003 100174701 BE0003565737 87,85 sell 40000 11/08/2006 100174701 BE0003565737 87,85 sell 3800 21/09/2006 100174701 BE0003565737 84,75 buy 530 23/11/2006 100174701 BE0003565737 85,3 buy 30 4/12/2006 100174701 BE0003565737 85,95 buy 20 23/01/2007 100174701 BE0003565737 97,65 buy 260 13/02/2007 100174701 BE0003565737 97,8 sell 400 23/02/2007 100174701 BE0003565737 96,2 buy 150 6/03/2007 100174701 BE0003565737 89,67 sell 300 6/03/2007 100174701 BE0003565737 89,67 sell 40000 ; run; proc sort data=i ; by tdate ; run; data i; set i; if _n_=1 then delete; run; data buy sell(rename=(tdate=_tdate fund=_fund isin=_isin price=_price transmod=_transmod numberstocks=_numberstocks)); set i; if transmod eq 'buy' then output buy; else if transmod eq 'sell' then output sell; run; data want; b=1;s=1; set buy nobs=b_obs; set sell nobs=s_obs; sum+numberstocks; _sum+_numberstocks;b_flag=0;s_flag=1; output; do while(b lt b_obs or s lt s_obs); do while(sum lt _sum); b+1; set buy point=b; sum+numberstocks;b_flag=0; s_flag=1; output; if b ge b_obs and s ge s_obs then stop; end; do while(sum gt _sum); s+1; set sell point=s; _sum+_numberstocks;b_flag=1; s_flag=0; output; if b ge b_obs and s ge s_obs then stop; end; end; run; data op; set want; dif_sum=_sum - sum; run; data opp; set op; lag_dif_sum=abs(lag(dif_sum)); drop sum _sum dif_sum ; run; data temp; set opp; if _n_ eq 1 then _numberstocks = numberstocks; else do; if b_flag =0 and s_flag=1 then _numberstocks=lag_dif_sum; else if b_flag=1 and s_flag=0 then numberstocks=lag_dif_sum; end; run; data result; set temp; if b_flag=1 and s_flag=0 and numberstocks lt _numberstocks then _numberstocks =numberstocks; drop b_flag s_flag lag_dif_sum; run; proc print noobs;run; [/pre] Ksharp ... View more

Oh My God! It is the most complicated logic I have ever meeted. After works for half a day.Only can get partly outcome. Hope this will help you a bit. I will promote this code after. Actually I am very tired ,need some rest. [pre] data i; input tdate : DDMMYY10. fund : $10. ISIN : $20. Price : commaX8. transmod $ numberstocks; format tdate DDMMYY10. price commaX8.2; datalines; 19/02/2004 100174701 BE0003565737 47 buy 3000 19/03/2004 100174701 BE0003565737 44,8 buy 3800 24/03/2004 100174701 BE0003565737 44,59 buy 12050 23/06/2005 100174701 BE0003565737 66,65 buy 3900 11/11/2005 100174701 BE0003565737 73,15 buy 3000 14/11/2005 100174701 BE0003565737 73,2 buy 3800 17/05/2006 100174701 BE0003565737 84,3 buy 250 12/06/2006 100174701 BE0003565737 78,05 buy 800 11/08/2003 100174701 BE0003565737 87,85 sell 40000 11/08/2006 100174701 BE0003565737 87,85 sell 3800 21/09/2006 100174701 BE0003565737 84,75 buy 530 23/11/2006 100174701 BE0003565737 85,3 buy 30 4/12/2006 100174701 BE0003565737 85,95 buy 20 23/01/2007 100174701 BE0003565737 97,65 buy 260 13/02/2007 100174701 BE0003565737 97,8 sell 400 23/02/2007 100174701 BE0003565737 96,2 buy 150 6/03/2007 100174701 BE0003565737 89,67 sell 300 6/03/2007 100174701 BE0003565737 89,67 sell 40000 ; run; proc sort data=i ; by tdate ; run; data i; set i; if _n_=1 then delete; run; data buy sell(rename=(tdate=_tdate fund=_fund isin=_isin price=_price transmod=_transmod numberstocks=_numberstocks)); set i; if transmod eq 'buy' then output buy; else if transmod eq 'sell' then output sell; run; data want; b=1;s=1; sum=0;_sum=0; set buy nobs=b_obs;set sell nobs=s_obs; do i=1 to 100000; *10000 is whatever you want ,just to be enough bigger than number of obs.; if numberstocks lt _numberstocks then do; do until(_numberstocks lt sum); set buy point=b ; b+1; output; sum=sum(sum,numberstocks); end; s+1; _numberstocks=_numberstocks - ( sum - numberstocks ); sum=0; put _numberstocks= numberstocks= sum= b= s= b_obs= s_obs=; if b ge b_obs then stop; *if s ge s_obs then stop; end; else if numberstocks gt _numberstocks then do; do until( numberstocks lt _sum ); set sell point=s ; s+1; output; _sum=sum(_sum,_numberstocks); end; b+1; numberstocks=numberstocks - ( _sum - _numberstocks ); _sum=0; put _numberstocks= numberstocks= _sum= b= s= b_obs= s_obs=; if s ge s_obs then stop; *if b ge b_obs then stop; end; end; run; data op; set want; range=_numberstocks - numberstocks; cul_sum+range; dif=range-cul_sum; run; [/pre] Ksharp ... View more

Hi. In SAS ,there are four input method : list input, formatted input, column input , named input. The difference you refer to is between list input and formatted input.Art has some details for it. So If you understand these four input way,then will process complicated data perfently. Ksharp Message was edited by: Ksharp ... View more

Hi. Cynthia ,I think Function COMPRESS() is more powerful and useful than what you refer to. [pre] data temp; infile datalines truncover; input address $100.; datalines; 1234 Sesame St. 98765 Chocolate Way North 1234 Grouch Lane 98765-A 34th St, South ; run; data address; set temp; only_digit=compress(address,,'ap'); only_char_punctuation=compress(address,,'d'); run; proc print;run; [/pre] Ksharp Message was edited by: Ksharp ... View more