I have an example of PROC SURVEYSELECT where I created four groups containing five IDs in each group. I want to be able to take a random sample where the IDs in different stratifications (i.e. groups) do not overlap. How can I accomplish this? Note that each group has the same repeating ID - 1 and 2. The next three IDs are unique to the group.
Example code:
data survey;
input group $ id;
datalines;
a 1
a 2
a 3
a 4
a 5
b 1
b 2
b 6
b 7
b 8
c 1
c 2
c 9
c 10
c 11
d 1
d 2
d 12
d 13
d 14
;
proc surveyselect data=survey
method=srs n=3
out=MyStratExample;
strata group;
run;
proc print data=MyStratExample;
run;
current output:
a 1 0.6 1.6666666667
a 3 0.6 1.6666666667
a 4 0.6 1.6666666667
b 1 0.6 1.6666666667
b 2 0.6 1.6666666667
b 7 0.6 1.6666666667
c 1 0.6 1.6666666667
c 2 0.6 1.6666666667
c 11 0.6 1.6666666667
d 1 0.6 1.6666666667
d 2 0.6 1.6666666667
d 13 0.6 1.6666666667
We can observe that across the multiple groups SAS is taking samples of the same ID variable.
Related
I would like to plot dataset and obtain desired output with the right setup.
Plot the scatter such that the points are in shade red-color, from light red to dark red depending on the scale (ratio) of 0-1 (0=light red, 1=dark red).
Show the legend also showing the scale red color according to the ration 0-1 (point 1.)
Data explanation:
area - city (shortcut)
id - user id
var - variable
time - datetime
exit - consumer left
ratio - proportion (between 0-1)
Data sample and attempt plotting (obviously not correct):
data data;
input area $ id $ var $ time $ exit $ ratio $;
datalines;
A 1 1 1 0 0.18
A 1 1 2 0 0.11
A 2 1 1 1 0.14
A 2 1 2 0 0.15
A 2 1 3 0 0.14
A 3 1 1 0 0.17
A 3 1 2 0 0.19
A 3 1 3 1 0.21
A 3 1 4 0 0.14
B 4 2 1 0 0.14
B 4 2 2 1 0.15
B 5 2 1 0 0.17
B 5 2 2 0 0.25
B 5 2 3 0 0.31
A 1 3 1 0 0.22
A 1 3 2 0 0.13
A 2 3 1 1 0.16
A 2 3 2 0 0.11
A 2 3 3 0 0.22
A 3 3 1 0 0.27
A 3 3 2 0 0.29
A 3 3 3 1 0.31
A 3 3 4 0 0.24
B 4 4 1 0 0.24
B 4 4 2 1 0.35
B 5 4 1 0 0.47
B 5 4 2 0 0.15
B 5 4 3 0 0.21
;;
run;
data attrs;
input id $ risk $ fillcolor $;
datalines;
ratio 0.05 Verylightred
ratio 0.15 Lightred
ratio 0.20 Red
ratio 0.25 Darkred
ratio 0.30 Verydarkred
ratio 0.35 Verydarkstrongred
;
run;
proc sgpanel data=data dattrmap=attrs;
panelby area exit;
scatter y=id x=var / markerattrs = (symbol = squarefilled) group=ratio attrid=ratio;
run;
This will get you closer.
Ratio should be numeric to be graphed
Ratio is continuous, how should it be used to group?
For the colour on the data attribute map, the length of the colours is not long enough and risk should be numeric
I don't know exactly how to specify the ranges you'd like for the colours you'd like but this gets you closer using the automatic legend.
One way to get at this is to add the variable to the data set for each group and then you can control the colour of each group with the data attribute map. This would mean adding a column in the 'data' data set called ratio_group whcih maps to the values in the data attribute map table. Use that variable the group.
data data;
input area $ id $ var $ time $ exit $ ratio ;
datalines;
A 1 1 1 0 0.18
A 1 1 2 0 0.11
A 2 1 1 1 0.14
A 2 1 2 0 0.15
A 2 1 3 0 0.14
A 3 1 1 0 0.17
A 3 1 2 0 0.19
A 3 1 3 1 0.21
A 3 1 4 0 0.14
B 4 2 1 0 0.14
B 4 2 2 1 0.15
B 5 2 1 0 0.17
B 5 2 2 0 0.25
B 5 2 3 0 0.31
A 1 3 1 0 0.22
A 1 3 2 0 0.13
A 2 3 1 1 0.16
A 2 3 2 0 0.11
A 2 3 3 0 0.22
A 3 3 1 0 0.27
A 3 3 2 0 0.29
A 3 3 3 1 0.31
A 3 3 4 0 0.24
B 4 4 1 0 0.24
B 4 4 2 1 0.35
B 5 4 1 0 0.47
B 5 4 2 0 0.15
B 5 4 3 0 0.21
;;
run;
proc sgpanel data=data ;
panelby area exit;
scatter y=id x=var / markerattrs = (symbol = squarefilled size=10)
colorresponse=ratio
colormodel=(verylightred lightred red darkred verydarkred verydarkstrongred);
colaxis grid minorgrid;
rowaxis grid minorgrid;
run;
For marker size look at the SIZE option under the MARKERATTRS option.
For grids, look at the GRID/MINORGRID options under the COLAXIS and ROWAXIS statements.
COLAXIS documentation
In this data, I need to subset by each variable by certain percentage.
For example,
Obs Group Score
1 A 1
2 A 2
3 B 1
4 B 1
5 C 3
6 C 1
7 C 1
8 A 1
9 A 3
10 A 1
11 A 2
12 B 3
13 C 2
I would need to subset 10 obs.
The sample must consist of all groups, and score of 1 takes higher priority.
Each group is given certain percent.
Let say 50% for A, 20% for B and 30% for C.
I tried using proc surveyselect but it failed. The number of alloc is not same as the strata.
proc surveyselect data=example out=test sampsize=10;
strata group score/alloc=(0.5 0.2 0.3);
run;
I don't know proc surveyselect too much, so I give the data step version.
data have;
input Obs Group$ Score;
cards;
1 A 1
2 A 2
3 B 1
4 B 1
5 C 3
6 C 1
7 C 1
8 A 1
9 A 3
10 A 1
11 A 2
12 B 3
13 C 2
;
run;
proc sort;
by Group Score;
run;
data want;
array _Dist_[3]$ _temporary_('A','B','C');
array _Upper_[3] _temporary_(5,2,3);
array _Count_[3] _temporary_;
do i = 1 to rec;
set have nobs=rec point=i;
do j = 1 to dim(_Dist_);
_Count_[j] + (Group=_Dist_[j]);
if _Count_[j] <= _Upper_[j] and Group = _Dist_[j] then output;
end;
end;
stop;
drop j;
run;
I have data which is as follows:
data have;
length
group 8
replicate $ 1
day 8
observation 8
;
input (_all_) (:);
datalines;
1 A 1 0
1 A 1 5
1 A 1 3
1 A 1 3
1 A 2 7
1 A 2 2
1 A 2 4
1 A 2 2
1 B 1 1
1 B 1 3
1 B 1 8
1 B 1 0
1 B 2 3
1 B 2 8
1 B 2 1
1 B 2 3
1 C 1 1
1 C 1 5
1 C 1 2
1 C 1 7
1 C 2 2
1 C 2 1
1 C 2 4
1 C 2 1
2 A 1 7
2 A 1 5
2 A 1 3
2 A 1 1
2 A 2 0
2 A 2 5
2 A 2 3
2 A 2 0
2 B 1 0
2 B 1 3
2 B 1 4
2 B 1 8
2 B 2 1
2 B 2 3
2 B 2 4
2 B 2 0
2 C 1 0
2 C 1 4
2 C 1 3
2 C 1 1
2 C 2 2
2 C 2 3
2 C 2 0
2 C 2 1
3 A 1 4
3 A 1 5
3 A 1 6
3 A 1 7
3 A 2 3
3 A 2 1
3 A 2 5
3 A 2 2
3 B 1 2
3 B 1 0
3 B 1 2
3 B 1 3
3 B 2 0
3 B 2 6
3 B 2 3
3 B 2 7
3 C 1 7
3 C 1 5
3 C 1 3
3 C 1 1
3 C 2 0
3 C 2 3
3 C 2 2
3 C 2 1
;
run;
I want to split observation into two columns based on day.
observation_ observation_
Obs group replicate day_1 day_2
1 1 A 0 7
2 1 A 5 2
3 1 A 3 4
4 1 A 3 2
5 1 B 1 3
6 1 B 3 8
7 1 B 8 1
8 1 B 0 3
9 1 C 1 2
10 1 C 5 1
11 1 C 2 4
12 1 C 7 1
13 2 A 7 0
14 2 A 5 5
15 2 A 3 3
16 2 A 1 0
17 2 B 0 1
18 2 B 3 3
19 2 B 4 4
20 2 B 8 0
21 2 C 0 2
22 2 C 4 3
23 2 C 3 0
24 2 C 1 1
25 3 A 4 3
26 3 A 5 1
27 3 A 6 5
28 3 A 7 2
29 3 B 2 0
30 3 B 0 6
31 3 B 2 3
32 3 B 3 7
33 3 C 7 0
34 3 C 5 3
35 3 C 3 2
36 3 C 1 1
The observant SO reader will notice that I have asked essentially the same question previously. However, because of SAS's obsession with "levels" and "by groups", since the variable being used to split the variable of interest isn't binary, that solution doesn't generalize.
Trying it directly, the following occurs:
proc sort data = have out = sorted;
by
group
replicate
;
run;
proc transpose data = sorted out = test;
by
group
replicate
;
var observation;
id day;
run;
ERROR: The ID value "_1" occurs twice in the same BY group.
I can use a LET statement to repress the errors, but in addition to cluttering up the log, SAS retains only the last observation of each BY group.
proc sort data = have out = sorted;
by
group
replicate
;
run;
proc transpose data = sorted out = test let;
by
group
replicate
;
var observation;
id day;
run;
Obs group replicate _NAME_ _1 _2
1 1 A observation 3 2
2 1 B observation 0 3
3 1 C observation 7 1
4 2 A observation 1 0
5 2 B observation 8 0
6 2 C observation 1 1
7 3 A observation 7 2
8 3 B observation 3 7
9 3 C observation 1 1
I don't doubt there's some kludgy way it could be done, such as splitting each group into a separate data set and then re-merging them. It seems like it should be doable with PROC TRANSPOSE, although how escapes me. Any ideas?
Not sure what you're talking about with "SAS's obsession...", but the issue here is fairly straightforward; you need to tell SAS about the four rows (or whatever) being separate, distinct rows. by tells SAS what the row-level ID is, but you're lying to it when you say by group replicate, since there are still multiple rows under that. So you need to have a unique key. (This would be true in any database-like language, nothing unique to SAS here. )
I would do this - make a day_row field, then sort by that.
data have_id;
set have;
by group replicate day;
if first.day then day_row = 0;
day_row+1;
run;
proc sort data=have_id;
by group replicate day_row;
run;
proc transpose data=have_id out=want(drop=_name_) prefix=observation_day_;
by group replicate day_row;
var observation;
id day;
run;
Your output looks like you don't want to transpose the data but instead just want split it into DAY1 and DAY2 sets and merge them back together. This will just pair the multiple readings per BY group in the same order that they appear, which is what it looks like you did in your example.
data want ;
merge
have(where=(day=1) rename=(observation=day_1))
have(where=(day=2) rename=(observation=day_2))
;
by group replicate;
drop day ;
run;
You can read the source data as many times as you need for the number of values of DAY.
If you think that you might not have the same number of observations per BY group for each DAY then you should add these statements at the end of the data step.
output;
call missing(of day_:);
I have a dataset that looks like this
data test;
input id1$ id2$ score1 score2 score3 total;
datalines;
A D 9 36 6 51
A D 9 8 6 23
A E 5 3 2 10
B D 5 3 3 11
B E 7 4 7 18
B E 5 3 3 11
C D 8 7 9 24
C E 8 52 6 66
C D 4 5 3 12
;
run;
I want to add a column that calculates what percentage of the corresponding total is of the summation within id1 and id2.
What I mean is this; id1 has a value of A. Within the value of A, there are twoid2 values ; D and E. There are two values of D, and one of E. The two total values of D are 51 and 23, and they sum to 74. The one total value of E is 10, and it sums to 10. The column I'd like to create would hold the values of .68 (51/74), .31 (23/74), and 1 (10/10) in row 1 ,row 2, and row 3 respectively.
I need to perform this calculations for the rest of the id1 and their corresponding id2. So when complete, I want a table that would like like this:
id1 id2 score1 score2 score3 total percent_of_total
A D 9 36 6 51 0.689189189
A D 9 8 6 23 0.310810811
A E 5 3 2 10 1
B D 5 3 3 11 1
B E 7 4 7 18 0.620689655
B E 5 3 3 11 0.379310345
C D 8 7 9 24 0.666666667
C E 8 52 6 66 1
C D 4 5 3 12 0.333333333
I realize a loop might be able to solve the problem I've given, but I'm dealing with EIGHT levels of stratification, with as many as 98 sublevels within those levels. A loop is not practical. I'm thinking something along the lines of PROC SUMMARY but I'm not too familiar with the function.
Thank you.
It is easy to do with a data step. Make sure the records are sorted.
You can find the grand total for the ID1*ID2 combination and then use it to calculate the percentage.
proc sort data=test;
by id1 id2;
run;
data want ;
do until (last.id2);
set test ;
by id1 id2 ;
grand = sum(grand,total);
end;
do until (last.id2);
set test ;
by id1 id2 ;
precent_of_total = total/grand ;
output;
end;
run;
Suppose I've a dataset in the form:
A B C
1 3 5
1 4 8
1 3 3
2 2 2
2 7 6
2 3 3
3 4 4
3 4 7
3 2 8
Now, I want to take weighted average of each segment of A and then add them up over A. For example in A var for 1, I want to take the weighted avg as (3*5+4*8+3*3)/(3+4+3). And then add up to get 5.6. Same with other 2 segments of A. So, finally the table looks like the following:
A B C D
1 3 7 5.6
2 6 6 7
3 5 9 8.2
Thank you.
Just to provide an alternative approach, you can use the WEIGHT statement in PROC SUMMARY to achieve the same result. The only thing I'm not clear on from your example final table table is where the values of columns B & C come from (I've left these out of my solution below).
proc summary data=test nway;
class a;
var c / weight=b;
output out=agg2 (drop=_:) mean=d;
run;
You can find the solution below. I am curious about your result. For A=2, the weighted average should be (2*2+7*6+3*3)/(2+7+3), about 4.5. Why here you have 7?
data test;
input a b c ;
datalines;
1 3 5
1 4 8
1 3 3
2 2 2
2 7 6
2 3 3
3 4 4
3 4 7
3 2 8
;
run;
proc sql;
create table agg as
select a, b, c, sum(b*c)/sum(b) as d from test
group by a;
quit;
proc sort data=agg nodupkey;
by a d;
run;