About echoli

echoli · ‎06-19-2017

Hi All, I have a data that each patient has multiple records. how can I do: if the variable time has more than four records (include four) within any continuous 12 month for the patient, then the patient defined as "Serious"? Thanks, data abc; informat date mmddyy10.; informat time anydtdtm16.; input ID Date time; format date mmddyy10. time MDYAMPM.; datalines; 14022 6/22/2010 10/11/2004 9:02 14022 6/22/2010 3/25/2008 12:53 14022 6/22/2010 4/29/2008 9:04 14022 6/22/2010 6/17/2008 9:14 14022 6/22/2010 7/8/2008 10:41 14022 6/22/2010 8/12/2008 12:30 14022 6/22/2010 9/16/2008 14:24 14022 6/22/2010 10/21/2008 1:44 14022 6/22/2010 11/18/2008 2:11 14022 6/22/2010 12/30/2008 9:53 14022 6/22/2010 2/10/2009 13:07 14022 6/22/2010 4/7/2009 9:45 14022 6/22/2010 6/2/2009 7:35 14022 6/22/2010 7/28/2009 10:11 14022 6/22/2010 9/29/2009 10:46 14022 6/22/2010 12/1/2009 10:51 14022 6/22/2010 1/26/2010 8:49 14022 6/22/2010 5/23/2010 7:48 14022 6/22/2010 6/22/2010 7:46 14032 2/7/2012 7/24/2002 10:25 14032 2/7/2012 1/21/2003 12:00 14032 2/7/2012 1/28/2003 9:25 14032 2/7/2012 1/29/2008 9:00 14032 2/7/2012 3/20/2008 13:52 14032 2/7/2012 5/1/2008 8:53 14032 2/7/2012 6/12/2008 14:01 14032 2/7/2012 7/24/2008 15:00 14032 2/7/2012 9/11/2008 11:02 14032 2/7/2012 9/20/2011 9:25 14032 2/7/2012 11/1/2011 10:40 14032 2/7/2012 1/8/2012 12:58 14032 2/7/2012 2/7/2012 10:20 14032 2/7/2012 3/7/2012 9:18 14032 2/7/2012 5/15/2012 12:43 14032 2/7/2012 6/19/2012 9:19 14032 2/7/2012 7/31/2012 11:48 14032 2/7/2012 9/18/2012 12:15 14032 2/7/2012 10/16/2012 12:35 14032 2/7/2012 10/18/2012 11:12 14032 2/7/2012 10/19/2012 5:00 14032 2/7/2012 10/20/2012 4:33 14032 2/7/2012 10/21/2012 4:19 14032 2/7/2012 10/22/2012 4:30 14032 2/7/2012 10/23/2012 5:03 14032 2/7/2012 10/24/2012 3:14 14032 2/7/2012 10/30/2012 11:10 14032 2/7/2012 11/27/2012 12:30 14032 2/7/2012 1/8/2013 9:58 14032 2/7/2012 2/19/2013 10:07 14032 2/7/2012 4/2/2013 13:45 14032 2/7/2012 5/14/2013 10:18 14601 1/30/2014 4/26/2005 8:40 14601 1/30/2014 6/28/2005 9:05 14601 1/30/2014 7/26/2005 8:21 14601 1/30/2014 8/9/2005 8:24 14601 1/30/2014 8/23/2005 8:10 14601 1/30/2014 10/18/2005 9:20 14601 1/30/2014 12/29/2013 9:32 14601 1/30/2014 12/30/2013 9:04 14601 1/30/2014 2/16/2014 8:44 14601 1/30/2014 5/2/2006 10:31 14601 1/30/2014 7/12/2006 7:31 14601 1/30/2014 9/7/2006 8:55 14601 1/30/2014 10/31/2006 8:43 14601 1/30/2014 2/6/2007 9:15 14601 1/30/2014 5/3/2007 7:54 14601 1/30/2014 6/28/2007 7:52 14601 1/30/2014 8/23/2007 8:10 14601 1/30/2014 10/18/2007 8:17 14601 1/30/2014 11/19/2007 2:07 14601 1/30/2014 12/13/2007 8:16 14601 1/30/2014 1/24/2008 10:23 14601 1/30/2014 1/31/2008 10:07 14601 1/30/2014 2/12/2008 12:31 14601 1/30/2014 3/25/2008 9:52 ; run;

echoli · ‎06-19-2017

Hi art297, I tried to use the slolution you helped me, it works!! but when I use my real data (I have lots of columns and rows) it doesn't work. it shows " The query requires remerging summary statistics back with the original data." in the log. I doubt. Do you know where is the problem? Thanks, L 1455 proc sql noprint; 1456 create table want as 1457 select mrn, event_end_UA, event_end_dt 1458 from combine_CBC 1459 where intnx('month', event_end_UA, -1, 's') <= datepart(event_end_dt) <= 1459! intnx('month', event_end_UA, 1, 's') 1460 group by mrn, event_end_UA 1461 having abs(datepart(event_end_dt) - event_end_UA) = min(abs(datepart(event_end_UA)- 1461! event_end_dt)) 1462 ; NOTE: The query requires remerging summary statistics back with the original data. NOTE: Table WORK.WANT created, with 0 rows and 3 columns. 1463 quit; NOTE: PROCEDURE SQL used (Total process time): real time 0.03 seconds cpu time 0.04 seconds

echoli · ‎06-19-2017

Thanks for your explanation, Ballardw! That's really my goal!!

echoli · ‎06-16-2017

Got it!!! Thanks for your help, Astounding!!

echoli · ‎06-16-2017

Hi, Here is a dataset, how can I select rows if the event_time closest to date measured within one month. For example, for id = 14022, the date is 6/22/2010, I want to select reows that event_time should be in 5/22/2010 to 7/22/2010. (if there are two measures within one month, choose the closest one. if there is no measure within one month, then NA.). Any clue or idea? data abc; input @1 ID @8 Date mmddyy10. @20 time anydtdtm.; format date mmddyy10. time MDYAMPM.; datalines; 14022 6/22/2010 10/11/2004 9:02 14022 6/22/2010 3/25/2008 12:53 14022 6/22/2010 4/29/2008 9:04 14022 6/22/2010 6/17/2008 9:14 14022 6/22/2010 7/8/2008 10:41 14022 6/22/2010 8/12/2008 12:30 14022 6/22/2010 9/16/2008 14:24 14022 6/22/2010 10/21/2008 1:44 14022 6/22/2010 11/18/2008 2:11 14022 6/22/2010 12/30/2008 9:53 14022 6/22/2010 2/10/2009 13:07 14022 6/22/2010 4/7/2009 9:45 14022 6/22/2010 6/2/2009 7:35 14022 6/22/2010 7/28/2009 10:11 14022 6/22/2010 9/29/2009 10:46 14022 6/22/2010 12/1/2009 10:51 14022 6/22/2010 1/26/2010 8:49 14022 6/22/2010 4/13/2010 7:48 14022 6/22/2010 6/22/2010 7:46 14032 2/7/2012 12/24/2002 10:25 14032 2/7/2012 1/21/2003 12:00 14032 2/7/2012 1/28/2003 9:25 14032 2/7/2012 1/29/2008 9:00 14032 2/7/2012 3/20/2008 13:52 14032 2/7/2012 5/1/2008 8:53 14032 2/7/2012 6/12/2008 14:01 14032 2/7/2012 7/24/2008 15:00 14032 2/7/2012 9/11/2008 11:02 14032 2/7/2012 9/20/2011 9:25 14032 2/7/2012 11/1/2011 10:40 14032 2/7/2012 12/22/2011 12:58 14032 2/7/2012 2/7/2012 10:20 14032 2/7/2012 3/20/2012 9:18 14032 2/7/2012 5/15/2012 12:43 14032 2/7/2012 6/19/2012 9:19 14032 2/7/2012 7/31/2012 11:48 14032 2/7/2012 9/18/2012 12:15 14032 2/7/2012 10/16/2012 12:35 14032 2/7/2012 10/18/2012 11:12 14032 2/7/2012 10/19/2012 5:00 14032 2/7/2012 10/20/2012 4:33 14032 2/7/2012 10/21/2012 4:19 14032 2/7/2012 10/22/2012 4:30 14032 2/7/2012 10/23/2012 5:03 14032 2/7/2012 10/24/2012 3:14 14032 2/7/2012 10/30/2012 11:10 14032 2/7/2012 11/27/2012 12:30 14032 2/7/2012 1/8/2013 9:58 14032 2/7/2012 2/19/2013 10:07 14032 2/7/2012 4/2/2013 13:45 14032 2/7/2012 5/14/2013 10:18 14601 1/30/2014 4/26/2005 8:40 14601 1/30/2014 6/28/2005 9:05 14601 1/30/2014 7/26/2005 8:21 14601 1/30/2014 8/9/2005 8:24 14601 1/30/2014 8/23/2005 8:10 14601 1/30/2014 10/18/2005 9:20 14601 1/30/2014 12/13/2005 9:32 14601 1/30/2014 2/7/2006 9:04 14601 1/30/2014 2/16/2006 8:44 14601 1/30/2014 5/2/2006 10:31 14601 1/30/2014 7/12/2006 7:31 14601 1/30/2014 9/7/2006 8:55 14601 1/30/2014 10/31/2006 8:43 14601 1/30/2014 2/6/2007 9:15 14601 1/30/2014 5/3/2007 7:54 14601 1/30/2014 6/28/2007 7:52 14601 1/30/2014 8/23/2007 8:10 14601 1/30/2014 10/18/2007 8:17 14601 1/30/2014 11/19/2007 2:07 14601 1/30/2014 12/13/2007 8:16 14601 1/30/2014 1/24/2008 10:23 14601 1/30/2014 1/31/2008 10:07 14601 1/30/2014 2/12/2008 12:31 14601 1/30/2014 3/25/2008 9:52 ; run; Thanks,

echoli · ‎06-16-2017

Got it! Thanks Astounding! Can I have one more question: what is the role for already_output = 'Y' in the code? Thanks, L

echoli · ‎06-16-2017

Hi Astounding, May I ask a question? in the code you answer me, what is the role for retain statement? we've already create highest_result, highest_date, already_output variables, in my opinion, i can run outcome out without retain statement, but the outcome is wrong. So, what is the role of retain statement in the code? Thanks, L

echoli · ‎06-16-2017

Thanks so much, Astounding. That's exactly what I want.

echoli · ‎06-15-2017

Sorry about that. I've edit the message. Thanks,

echoli · ‎06-15-2017

Hi All, I have a data below. I want to select rows meet criterias: 1. if the ID is unique (just one record) then retain it. 2. if the ID has multiple records, after proc sort data, choose the first record that result >= 30. 3. if the ID has multiple records, after proc sort data, if all the results for this ID <30 then retain the largest value. I can do 1 and 3 criteria, but no idea about 2. Any one can help? data have; input ID Date mmddyy10. Result; datalines; 14022 6/22/2010 6 14032 6/28/2011 5 14032 2/7/2012 8 14102 4/13/2010 6 14102 6/24/2009 16 14129 11/8/2011 9 14129 12/29/2009 24 14209 9/29/2011 10 14209 9/24/2013 12 14238 8/18/2011 4 14273 5/31/2011 18 14399 1/21/2014 4 14399 4/10/2012 10 14424 11/6/2012 10 14593 11/8/2013 20 14593 5/31/2011 22 14593 1/7/2014 26 14593 9/13/2011 46 14601 8/29/2013 2 14601 2/29/2012 3 14601 1/30/2014 4 14601 11/15/2011 5 14601 7/21/2009 6 14601 8/24/2009 13 14621 11/10/2009 4 14621 1/24/2012 7 14726 11/20/2012 11 14726 10/14/2014 13 14726 3/27/2012 20 14868 10/11/2012 7 14868 5/28/2009 10 15597 7/3/2012 5 15597 8/24/2010 5 17246 6/10/2014 8 17246 10/9/2012 16 17638 6/11/2013 3 17638 1/3/2013 3 17672 5/17/2011 3 17950 5/14/2013 4 17950 3/13/2012 4 17950 4/20/2010 6 17950 5/3/2010 7 17955 12/11/2012 8 17955 10/1/2013 12 17955 4/19/2011 13 17955 11/11/2014 26 ; Thanks

echoli · ‎06-14-2017

Hi All, I have a dataset below: data want; input Group $ Timepoint GM FRO CONT RES; datalines; A 1 3 2 5 2 A 2 2 3 3 8 A 1 4 3 5 6 A 2 5 4 3 7 B 1 6 4 3 5 B 2 5 4 3 6 B 1 4 5 2 7 ; How can I add 5 for GM - RES if timepoint = 1? Thanks, L

echoli · ‎06-08-2017

Hi Ballardw, Thanks so much for your help. That works!!! I'll pay attention to dataset next time. C

echoli · ‎06-08-2017

Hi, I have a dataset below, I want to seperate data into four groups by Group (TX, Cont) and Timepoint (1, 2). here is my code: data want1; set want; if timepoint = 1 then if group = "TX" then do result1_1 = result1; result2_1 = result2; else result1_2 = result1; result2_2 = result2; end; else if timepoint = 2 then if group = "TX" then do result1_3 = result1; result2_3 = result2; else result1_4= result1; result2_4 = result2; output; end; run; is my code wrong? the ouput is not what I want. Here is my data: Group Timepoint Result1 result2 A 1 3 2 A 2 2 3 A 1 4 3 A 2 5 4 A 1 6 4 A 2 5 4 A 1 4 5 A 2 7 5 A 1 23 5 A 2 6 5 B 1 5 5 B 2 8 6 B 1 6 6 B 2 7 6 B 1 456 6 B 2 56 7 B 1 655 7 B 2 9 8 B 1 5 8 B 2 6 9 B 1 5 23 B 2 3 56 B 1 8 456 B 2 4 655 I want my result like below: Result1_1 Result2_1 Result1_2 Result2_2 Result1_3 Result2_3 Result1_4 Result2_4 3 2 2 3 5 5 8 6 4 3 5 4 6 6 7 6 6 4 5 4 456 6 56 7 4 5 7 5 655 7 9 8 23 5 6 5 5 8 6 9 5 23 3 56 8 456 4 655 Thanks,

echoli · ‎06-07-2017

Hi Rick, In my picture, there are four kinds of data points, matched with 1, 2, 3, 4 you helped me list in the reply. I want four colors for these four kinds of data points, and also four lines matched with same colors. Like in the vector statement, I want X, Y, Xorigin, Yorigin have different colors, is that possible? Also, the four lines create by X, Y, xorigin, yorigin columns in Loess statement should match same colors as vector statement. Is my explain clear? Thanks so much for your patience to help me! C

echoli · ‎06-07-2017

Hi Rick, Yes, 1,2,3,4 are the four lines I want. I want four lines with four different colors. In my picture, 1 and 2 have the same color, 3 and 4 have the same color. That's what I want to solve. Thanks, C

Online Status	Offline
Date Last Visited	‎10-30-2017 03:59 PM

Re: points not on the same vertical line

Re: points not on the same vertical line

Re: points not on the same vertical line

points not on the same vertical line

Re: data symbles doesn't shown in sgplot

data symbles doesn't shown in sgplot

Re: if with multiple conditions

if with multiple conditions across observations

Re: dataset merge

Re: dataset merge

Re: add a value in multiple columns

Re: add a value in multiple columns

Re: add a value in multiple columns

change rows order

identify observations in rolling 12 month period

Re: select rows within one month of a time

Re: select rows within one month of a time

Re: select row in multiple rows by several conditions

select rows within one month of a time

Re: select row in multiple rows by several conditions

Re: select row in multiple rows by several conditions

Re: select row in multiple rows by several conditions

Re: select row in multiple rows by several conditions

select row in multiple rows by several conditions

add a value in multiple columns

Re: manage data into four groups

manage data into four groups

Re: how to change line colors with different groups and different poin...

Re: how to change line colors with different groups and different poin...