About monday89

monday89 · ‎02-09-2020

I have this following code to export . I would like to bold all columns but won't work ods tagsets.simplelatex file=".FILE DESITINATION \filename.tex" (notop nobot) stylesheet="sas.sty" (url="sas"); options linesize=256 ; proc report data=data nowindows headline headskip style(header)={font_weight=bold} style(report)={outputwidth=75%}; column column_1 Frequency Percent; define column_1 /display "column 1" width = 14 ; define Frequency / display "Frequency" width =14; define Percent/ display "Percent" order descending; run; ods tagsets.simplelatex close;

monday89 · ‎02-09-2020

I have a following dataset. WHat I want to do is count TOTAL number of "term" per "class". Duplicates are OK. For example Patient ID = 1 has two "term 2" but with different dates so it should be counted twice. Total denominator should be grouped by "Class" denominator. But unique adult and baby should also be counted. Raw dataset treatment_group ID class term date group adult baby placebo 2 GI related term 1 1/3/2020 adult 1 0 placebo 2 GI related term 2 2/1/2020 adult 1 0 placebo 5 Heart related term 2 2/4/2020 baby 0 1 placebo 5 Heart related term 3 2/5/2020 baby 0 1 treatment 3 General disorders term 1 2/2/2019 adult 1 0 treatment 4 General disorders term 2 2/3/2020 adult 1 0 treatment 1 Heart related term 1 1/2/2019 baby 0 1 treatment 1 Heart related term 2 1/2/2019 baby 0 1 treatment 1 Heart related term 2 1/4/2019 baby 0 1 Want: treatment_group Class Term event total_class_denominator adult child placebo GI related term 1 1 2 1 0 placebo GI related term 2 1 2 1 0 placebo Heart related term 2 1 2 0 1 placebo Heart related term 3 1 2 0 1 treatment General disorders term 1 1 2 1 0 treatment General disorders term 2 1 2 1 0 treatment Heart related term 1 1 3 0 1 treatment Heart related term 2 2 3 0 1 I tried the following proc sql; select treatment_group, class, term, count(term) as total_event_numerator, count(class) as total_event_denominator, sum(age_child) as child, sum(age_infant) as infant from test group by treatment,class, term quit;

monday89 · ‎02-08-2020

I am having trouble calculating n and total event for multiple group Here is the table: treatment_group AgeGroup patientID group1 group2 treatment adult 1 ONE headache treatment adult 1 ONE fever treatment adult 1 ONE cough treatment child 2 ONE cough treatment child 3 TWO heart disease treatment child 3 TWO smoker treatment adult 4 TWO heart disease placebo adult 5 ONE headache placebo adult 5 ONE cough placebo child 6 ONE pneumonia placebo child 6 TWO heart disease placebo child 7 TWO heart disease placebo child 7 TWO smoker I want to get total number per group 2 per group 1 per treatment_group for total patients for each age group. Something like this: Total_event will be denominator for group1 based on per treatment_group and agegroup treatment_group AgeGroup group1 group2 event total_event placebo adult ONE cough 1 2 placebo adult ONE headache 1 2 placebo child ONE pneumonia 1 1 placebo child TWO heart disease 2 3 placebo child TWO smoker 1 3 treatment adult ONE cough 1 3 treatment adult ONE fever 1 3 treatment adult ONE headache 1 3 treatment adult TWO heart disease 1 1 treatment child ONE cough 1 1 treatment child TWO heart disease 1 2 treatment child TWO smoker 1 2 i tried but i haven't been successful for age group and treatment at the same time proc freq data=test ; table group1*group2/missing; by treatment_group ; run;

monday89 · ‎02-08-2020

I have three data sets 1. Patient Level - all unique patient with patient ID [table1] patientID 123-ABC 123-DEF 123-GHI 123-JKL 123-MNO 123-PQR 123-STU 2. Reference table - Reference name for each number [table2] category number ABC 1 DEF 2 PER 3 WER 4 QWE 5 3. Patient level having certain patient number [table3] patientID number 123-ABC 1 123-DEF 1 123-GHI 2 123-GHI 3 123-JKL 4 123-JKL 2 123-MNO 5 What I want is that combine all THREE tables and have 1 or 0 even though there is no outer join. The patient list should have 1 or 0 for each reference number when appropriate patientID category measure 123-ABC ABC 1 123-ABC DEF 0 123-ABC PER 0 123-ABC WER 0 123-ABC QWE 0 123-DEF ABC 1 123-DEF DEF 0 123-DEF PER 0 123-DEF WER 0 123-DEF QWE 0 123-GHI ABC 0 123-GHI DEF 1 123-GHI PER 1 123-GHI WER 0 123-GHI QWE 0 123-JKL ABC 0 123-JKL DEF 0 123-JKL PER 1 123-JKL WER 1 123-JKL QWE 0 123-MNO ABC 0 123-MNO DEF 0 123-MNO PER 0 123-MNO WER 0 123-MNO QWE 1 123-PQR ABC 0 123-PQR DEF 0 123-PQR PER 0 123-PQR WER 0 123-PQR QWE 0 123-STU ABC 0 123-STU DEF 0 123-STU PER 0 123-STU WER 0 123-STU QWE 0 I tried the following proc sql; select a.patientid, b.category, b.number from table1 a left join table3 b full outer join a.patientid = b.patientid; quit; i am not sure how to use inner join

monday89 · ‎02-08-2020

Following table I want to calculate category2 per category group ID category2 binary placebo 1 ABC-123 0 placebo 3 ABC-123 0 placebo 1 DEF-134 1 placebo 3 DEF-134 1 placebo 1 OPS-129 1 placebo 3 OPS-129 0 placebo 1 TFD-157 0 placebo 3 TFD-157 1 treatment 2 ABC-123 1 treatment 4 ABC-123 1 treatment 2 DEF-134 1 treatment 4 DEF-134 0 treatment 2 OPS-129 0 treatment 4 OPS-129 1 treatment 2 TFD-157 1 treatment 4 TFD-157 1 to something like this: group category binary denominator placebo ABC-123 0 2 placebo DEF-134 2 2 placebo OPS-129 1 2 placebo TFD-157 1 2 treatment ABC-123 2 2 treatment DEF-134 1 2 treatment OPS-129 1 2 treatment TFD-157 2 2 the denominator is treatment group. i tried the following but i can only get the numerator and denominator separately proc sql; create table yes_group as select distinct group, category, count(binary) as binary_yes from have where binary= 1 group by group, category; quit; and i have to repeat for denominator. is there a bette way from proc freq?

monday89 · ‎02-08-2020

I have a 1) reference table for medications and 2) patient level data for medications Reference table for medications med_id med_name 1 ABC 2 DEF 3 GHI 4 JKL 5 MNO 6 PQR 7 STU 8 VWX 9 YZ 2) Patient Level Data patientId med_name medication 123 ABC 1 123 GHI 1 123 JKL 1 123 PQR 1 456 DEF 1 456 GHI 1 What I want to do is that join both reference med and patient level but indicate the patients that are not on certain medications. patientId med_name medication 123 ABC 1 123 DEF 0 123 GHI 1 123 JKL 1 123 MNO 0 123 PQR 1 123 STU 0 123 VWX 0 123 YZ 0 456 ABC 0 456 DEF 1 456 GHI 1 456 JKL 0 456 MNO 0 456 PQR 0 456 STU 0 456 VWX 0 456 YZ 0 What I did was: proc sql; select a.*, b.med_name from patient left join medtable a on med_name = med_name;quit;

monday89 · ‎02-07-2020

Sorry I meant as in if group is Baseline or Daily (consequential) for the same patient AND all time points and dates match then baseline = 1

monday89 · ‎02-07-2020

Current table patientid Event type date timepoint1 timepoint2 timepoint3 timepoint4 1 Baseline 1/28/2020 1200 600 1200 1 Daily 1/28/2020 1211 1211 1 Daily 1/29/2020 1221 1000 1221 2 Baseline 11/21/2019 1011 1359 1200 1200 2 Daily 11/21/2019 1011 1359 1200 1200 2 Daily 11/21/2019 1643 1645 1645 1644 2 Daily 11/21/2019 2225 2225 2225 2225 3 Baseline 11/25/2019 1440 1440 1440 1440 3 Daily 11/25/2019 1520 1520 1520 1520 4 Baseline 11/11/2019 1259 1300 1301 1300 4 Daily 11/11/2019 1259 1300 1301 1300 4 Daily 11/11/2019 1845 1847 1845 1845 within each same patient group, i would like to create a flag if date equals to next date within each patient AND all time points match, then baseline = 1 or 0 patientid Event type date timepoint1 timepoint2 timepoint3 timepoint4 baseline 1 Baseline 1/28/2020 1200 600 1200 1 1 Daily 1/28/2020 1211 1211 0 1 Daily 1/29/2020 1221 1000 1221 0 2 Baseline 11/21/2019 1011 1359 1200 1200 1 2 Daily 11/21/2019 1011 1359 1200 1200 1 2 Daily 11/21/2019 1643 1645 1645 1644 0 2 Daily 11/21/2019 2225 2225 2225 2225 0 3 Baseline 11/25/2019 1440 1440 1440 1440 1 3 Daily 11/25/2019 1520 1520 1520 1520 0 4 Baseline 11/11/2019 1259 1300 1301 1300 1 4 Daily 11/11/2019 1259 1300 1301 1300 1 4 Daily 11/11/2019 1845 1847 1845 1845 0 I tried to divide daily and baseline tables into TWO different tables and left join by date and then timepoints. i am really stuck

monday89 · ‎02-07-2020

Hi I have the following table: group event Frequency Percent CumulativeFrequency a fever 2 20.00 2 a cold 3 30.00 5 a cough 1 10.00 6 a flu 4 40.00 10 b headache 2 20.00 2 b fatigue 2 20.00 4 b fever 1 10.00 5 b cold 1 10.00 6 What I would like to do is that: take max cumulativefrequency of the "group" and create a static column. Something like this group event Frequency Percent CumulativeFrequency Total a fever 2 20.00 2 10 a cold 3 30.00 5 10 a cough 1 10.00 6 10 a flu 4 40.00 10 10 b headache 2 20.00 2 6 b fatigue 2 20.00 4 6 b fever 1 10.00 5 6 b cold 1 10.00 6 6 i tried to proc sql; create table test c as select group, max(CumulativeFrequency) as total from have; group by group; quit; it works but i have to rejoin test and have. is there a way without creating additional dataset?

monday89 · ‎02-07-2020

Yes, easier-to-create table will be acceptable. I am sure it is going to be much better than mine!

monday89 · ‎02-07-2020

I have the following table below patientid trt gender race cad_history 1 group 1 F white 1 2 group 2 F white 1 3 group 1 M asian 0 4 group 1 F black 1 5 group 2 M asian 0 proc sort data=work.test; by trt; run; ods output onewayfreqs=work.test2; proc freq data=work.test; table gender race cad_history /missing; by trt ; run; ods listing; I get this: trt Table gender gender Frequency Percent Cumulative Frequency CumulativePercent race race cad_history cad_history group 1 Table gender F F 2 66.67 2 66.67 . group 1 Table gender M M 1 33.33 3 100 . group 1 Table race 1 33.33 1 33.33 asian asian . group 1 Table race 1 33.33 2 66.67 black black . group 1 Table race 1 33.33 3 100 white white . group 1 Table cad_history 1 33.33 1 33.33 0 0 group 1 Table cad_history 2 66.67 3 100 1 1 group 2 Table gender F F 1 50 1 50 . group 2 Table gender M M 1 50 2 100 . group 2 Table race 1 50 1 50 asian asian . group 2 Table race 1 50 2 100 white white . group 2 Table cad_history 1 50 1 50 0 0 group 2 Table cad_history 1 50 2 100 1 1 But then I have to polish it more to get this: trt group 1 group 1 F number (%) number (%) M number (%) number (%) white number (%) number (%) asian number (%) number (%) black number (%) number (%) cad history 1 number (%) number (%) cad history 0 number (%) number (%) Is there a simple way to do this? rather than writing additional commands? Thanks

monday89 · ‎02-05-2020

I have the following table: Name DATE actual 1111 1/2/2019 1 1111 1/2/2019 1 1111 1/4/2019 0 1111 1/5/2019 0 1111 1/6/2019 0 2222 1/4/2019 1 2222 1/5/2019 1 2222 1/6/2019 1 3333 1/6/2019 1 3333 1/10/2019 0 3333 1/10/2019 0 3333 1/13/2019 1 3333 1/14/2019 1 3333 1/15/2019 1 I want total counts per Name (if two dates are the same twice, count as twice) and then only 1 for actual (sum) NAME TOTAL_COUNTS ACTUAL_COUNTS 1111 5 2 2222 3 3 3333 6 4 I tried the nested loop but no luck proc sql; select DISTINCT a.NAME, count(a.DATE) as Total_counts, a.actual_counts from (select distinct a.NAME, count(a.actual) as actual_counts from sql.p_IE_ELIGIBILITY a where a.actual="1" group by a.NAME) group by a.NAME as a full join (select distinct a.NAME, count(*) as total from TEST) as b on a.NAME = b.NAME; quit;

monday89 · ‎01-27-2020

great thank you it works! But now i have another issue. Currently I have only 2 groups: treatment and placebo, but if I add the third arm to be arm_3. I don’t want the order to change desc or ascending. It should be treatment, placebo, arm_3. If I add /group in the define, it doesn’t retain the order Group value Treatment 112 Treatment 312 Treatment 423 Placebo 242 Placebo 343 Arm_3 324 Arm_3 324 I wrote the following: proc report data=test nowindows headline headskip; column group value; define group/ group "Treatment Group" width =14; define value/ group "value" width =14; run; I am getting this [I do not want to change the order of group] Group value Arm_3 324 324 Placebo 242 343 Treatment 112 312 423

monday89 · ‎01-27-2020

I have two datasets dataset #1: by group group_assigned denominator placebo 10 treatment 12 dataset #2: Demographics treatment placebo Age Gender Race Medications the treatment groups changes every data export so i want to set n=# for each group. For this export, something like this: Demographics treatment (n=12) placebo (n=10) Age Gender Race Medications I tried %macro but i think it is not working

monday89 · ‎01-27-2020

I have the following table and I would like to merge and center the group by treatment groups group category value treatment 1 56 treatment 2 54 treament 3 32 placebo 1 45 placebo 2 34 placebo 3 23 Like this: I tried: proc report data=test nowindows split='~' spanrows style(report)={font_size=9pt } style(header)={just=center font_weight=bold font_size=9pt background=CXE0E0E0} style(column)={font_size=9pt protectspecialchars=on}; column group category value; run; not sure what i am doing wrong

Online Status	Offline
Date Last Visited	‎12-16-2020 03:47 PM

assigning numbers to each group

Re: Merging 2 datasets

Merging 2 datasets

missing values - fill the new ones

Re: proc sgplot

proc sgplot

Latex table format

adding days to start date

Line Graph by Percentage and Frequency

Line graph by group and percentage

Re: Total by group

Re: time constraint overlap

latex and proc report bold header

total by treatment, class, term and age groups

proc freq owf

joining 3 tables into one

proc freq vs. sql

coming two tables with reference table

Re: remove duplicate by date and time

remove duplicate by date and time

max sum per group

Re: proc freq or some other option

proc freq or some other option

nested loops sql or dataset

Re: proc report merge and center

adding numbers at top

proc report merge and center