Hi Community,
Here I have a dataset as a Master table(AAA) where I already fetch some data from Master(AAA) to sub_master table(BBB).
(Table1)
data AAA;
input ID date_A ddmmyy10. grade;
cards;
111111 02/07/2013 1
111111 02/04/2014 1
111111 29/12/2014 .
111111 05/04/2014 .
100000 01/10/2012 .
100000 08/10/2015 2
100000 10/01/2017 5
100000 29/05/2018 2
100000 30/12/2019 .
;
run;
In my BBB table I already fetch some data from AAA table using some condition. Now the (Table 2) is look like
ID date_B GRD
111111 02/07/2013 1
111111 02/04/2014 1
100000 10/01/2017 5
100000 29/05/2018 2
Table 1 = AAA
Table 2 = BBB
Now i have to additionally insert some rows from AAA to BBB.
The condition is
First choose the max(date_B) and find the min(date_A) > max(date_B) and WHERE grade = '.'
(Which means 111111 29/12/2014 . should affect)
(Which means 100000 30/12/2019 . should affect)
How can I edit my query to get the result? Any help will be much appreciated!
First choose the max(date_B) and find the min(date_A) > max(date_B) and WHERE grade = '.'
(Which means 111111 29/12/2014 . should affect)
(Which means 100000 30/12/2019 . should affect)
are you sure this statement is true for ID 111111 results, would record dated 4-5-2014 be less than 12-29-2014?
I believe this code does what you are asking for correctly.
data Master;
input ID dt : ddmmyy10. grade;
cards;
111111 02/07/2013 1
111111 02/04/2014 1
111111 29/12/2014 .
111111 05/04/2014 .
100000 01/10/2012 .
100000 08/10/2015 .
100000 10/01/2017 5
100000 29/05/2018 1
100000 30/12/2019 .
111122 05/01/1922 .
111122 07/12/1932 .
111122 05/01/1933 .
111122 05/01/1935 .
111122 05/01/2018 1
111122 05/02/2018 .
;
run;
proc sort data=master;
by id dt;
run;
data sub_master;
input
ID dt : ddmmyy10. GRD;
cards;
111111 02/07/2013 1
111111 02/04/2014 1
100000 10/01/2017 5
100000 29/05/2018 2
111122 05/01/2018 1
;
proc sort data=sub_master;
by id dt;
run;
proc sql;
create table min_master as
select distinct a.id, max(a.dt), a.dt, a.grade as grd
from master as a,
sub_master as b
where a.grade = . and a.id = b.id and a.dt > b.dt
group by a.ID, b.id
having a.dt = min(a.dt) and b.dt = max(b.dt)
;
quit;
data want;
set sub_master
min_master(drop=_TEMG001);
format dt date10.;
run;
proc sort data=want;
by id dt;
run;
First choose the max(date_B) and find the min(date_A) > max(date_B) and WHERE grade = '.'
(Which means 111111 29/12/2014 . should affect)
(Which means 100000 30/12/2019 . should affect)
are you sure this statement is true for ID 111111 results, would record dated 4-5-2014 be less than 12-29-2014?
I believe this code does what you are asking for correctly.
data Master;
input ID dt : ddmmyy10. grade;
cards;
111111 02/07/2013 1
111111 02/04/2014 1
111111 29/12/2014 .
111111 05/04/2014 .
100000 01/10/2012 .
100000 08/10/2015 .
100000 10/01/2017 5
100000 29/05/2018 1
100000 30/12/2019 .
111122 05/01/1922 .
111122 07/12/1932 .
111122 05/01/1933 .
111122 05/01/1935 .
111122 05/01/2018 1
111122 05/02/2018 .
;
run;
proc sort data=master;
by id dt;
run;
data sub_master;
input
ID dt : ddmmyy10. GRD;
cards;
111111 02/07/2013 1
111111 02/04/2014 1
100000 10/01/2017 5
100000 29/05/2018 2
111122 05/01/2018 1
;
proc sort data=sub_master;
by id dt;
run;
proc sql;
create table min_master as
select distinct a.id, max(a.dt), a.dt, a.grade as grd
from master as a,
sub_master as b
where a.grade = . and a.id = b.id and a.dt > b.dt
group by a.ID, b.id
having a.dt = min(a.dt) and b.dt = max(b.dt)
;
quit;
data want;
set sub_master
min_master(drop=_TEMG001);
format dt date10.;
run;
proc sort data=want;
by id dt;
run;
Hi @VDD
Your code works well.
But I have a reverse task from the dataset and I tried but it endup up with partial output.
Task 1 (COMPLETED)
First choose the MAX(date_B) and find the MIN(date_A) > MAX(date_B) and WHERE grade = '.'
TASK 2 (PARTIAL OP)
----- choose the MIN(date_B) and find the MAX(date_A) < MIN(date_B) and WHERE grade = '.'
The example dataset which I mention in earlier post performs well with TASK2. But not works in my original dataset.
/***** TASK 1 **********/ /* You suggested*/ /** COMPLETED**/
Proc Sql;
create table Mater_min as
select distinct a.P_IDA,
max(a.Record_Date_A),a.Record_Date_A,a.Grade_A
from Master as a, SubMaster as b
where a.Grade_A = '' and a.P_IDA = b.P_IDB and a.Record_Date_A > b.Record_date_B
group by a.P_IDA,b.P_IDB
having a.Record_Date_A = min(a.Record_Date_A) and b.Record_Date_B = max(b.Record_date_B);
quit;
/******* TASK 2****/ /*What I edited for reverse method*/ /*Results PArtial Output*******/
Proc Sql;
create table Mater_max as
select distinct a.P_IDA,
min(a.Record_Date_A),a.Record_Date_A,a.Grade_A
from Master as a, SubMaster as b
where a.Grade_A = '' and a.P_IDA = b.P_IDB and a.Record_Date_A < b.Record_Date_B
group by a.P_IDA,b.P_IDB
having a.Record_Date_A = max(a.Record_Date_A) and b.Record_Date_B = min(b.Record_date_B);
quit;
HERE I ATTACH the original dataset Please gothrough it and let me know my error.
Thanks in Advance!
Save $250 on SAS Innovate and get a free advance copy of the new SAS For Dummies book! Use the code "SASforDummies" to register. Don't miss out, May 6-9, in Orlando, Florida.
Learn how use the CAT functions in SAS to join values from multiple variables into a single value.
Find more tutorials on the SAS Users YouTube channel.
Ready to level-up your skills? Choose your own adventure.