I just started using SAS and I'm trying to combine columns.
I've got table mainData
A1 A2 A3 A4
1 4 7 10
2 5 8 11
3 6 9 12
I want to create a new table rearrangedData
Type Value
A1 1
A1 2
A1 3
A2 4
A2 5
A2 6
A3 7
A3 8
A3 9
A4 10
A4 11
A4 12
There must be a simple solution to this I just can't figure this out. I'm thinking of writing do loop, but what if I don't know size of a table or amount of lines in a specific column. I can't figure how I would get such information in SAS.
This somewhat unusual transformation can be done via a transpose and some array logic:
data have;
input A1 A2 A3 A4;
cards;
1 4 7 10
2 5 8 11
3 6 9 12
;
run;
proc transpose data = have out = tr name=type prefix = r;
run;
data want;
set tr;
array r{*} r:;
do i = 1 to dim(r);
value = r[i];
output;
end;
drop i r:;
run;
Also, this preserves the original order without requiring a sort.
Make a dummy variable, then transpose data.
data have;
set have;
id=_n_;
run;
proc transpose data=have out=temp;
by id;
var A1-A4;
run;
proc sort data=temp out=want(rename=(_name_=type col1=value) drop=id);
by _name_;
run;
If you want to preserve the original order then you could use the POINT= option on the SET statement to loop over the data set once per variable (column).
So this data set will read the first observations just to get the variables defined. Then define the array VALUES so that we can use DIM(VALUES) to know how many columns. Then it uses the POINT= and NOBS= options on the SET statement to control the other loop. It uses the VNAME() function to find the name of the current variable in the array.
data want ;
set have ;
array values _numeric_;
do col=1 to dim(values);
length type $32 value 8;
type=vname(values(col));
do row=1 to nobs ;
set have point=row nobs=nobs ;
value=values(col);
output;
keep type value;
end;
end;
stop;
run;
Related
I need to aggregate about ten different vars on different groupings using Proc SQL;
Is there a way to achieve SUM () OVER ( [ partition_by_clause ] order_by_clause) in one sql query with different partition by clauses.
I've made an example here
data have;
infile cards;
input a b c d e f;
cards;
1 2 3 4 5
2 2 4 5 6
1 4 3 4 7
3 4 4 5 8
;
run;
proc sql;
create table want as
select *,
sum a over partiton by (b,c) as a1,
sum b over partiton by (c,d) as b1
sum c over partiton by (d,e) as c1
sum d over partiton by (a,c) as d1
from have
;
quit;
I don't want to wirte multiple sql queries and grouping on different vars and calculating one var in each step.
Hope that makes sense.
Proc SQL does not implement windowing functions and thus partition syntax therein as found in other SQL implementations. You can only do partition by with passthrough SQL to a connection that allows such syntax.
You could perform such a computation in DATA step using hashes.
data have;
infile cards;
input a b c d e ;
cards;
1 2 3 4 5
2 2 4 5 6
1 4 3 4 7
3 4 4 5 8
;
run;
data want;
if 0 then set have;
length a1 b1 c1 d1 8;
declare hash a1s();
a1s.defineKey('b', 'c');
a1s.defineData('a1');
a1s.defineDone();
declare hash b1s();
b1s.defineKey('c', 'd');
b1s.defineData('b1');
b1s.defineDone();
declare hash c1s();
c1s.defineKey('d', 'e');
c1s.defineData('c1');
c1s.defineDone();
declare hash d1s();
d1s.defineKey('a', 'c');
d1s.defineData('d1');
d1s.defineDone();
do while (not end);
set have end=end;
if a1s.find() = 0 then a1+a; else a1=a; a1s.replace();
if b1s.find() = 0 then b1+b; else b1=b; b1s.replace();
if c1s.find() = 0 then c1+c; else c1=c; c1s.replace();
if d1s.find() = 0 then d1+d; else d1=d; d1s.replace();
end;
do while (not last);
set have end=last;
a1s.find();
b1s.find();
c1s.find();
d1s.find();
output;
end;
format _numeric_ 4.;
stop;
run;
I am matching files base on IDs numbers. I need to format a data set with the IDs to be matched, so that the same ID number is not repeated in column a (because column b's ID is the surviving ID after the match is completed). My list of IDs has over 1 million observations, and the same ID may be repeated multiple times in either/both columns.
Here is an example of what I've got/need:
Sample Data
ID1 ID2
1 2
3 4
2 5
6 1
1 7
5 8
The surviving IDs would be:
2
4
5
error - 1 no longer exists
error - 1 no longer exists
8
WHAT I NEED
ID1 ID2
1 2
3 4
2 5
6 5
5 7
7 8
I am, probably very obviously, a SAS novice, but here is what I have tried, re-running over and over again because I have some IDs that are repeated upward of 50 times or more.
Proc sort data=Have;
by ID1;
run;
This sort makes the repeated ID1 values consecutive, so the I could use LAG to replace the destroyed ID1s with the surviving ID2 from the line above.
Data Want;
set Have;
by ID1;
lagID1=LAG(ID1);
lagID2=LAG(ID2);
If NOT first. ID1 THEN DO;
If ID1=lagID1 THEN ID1=lagID2;
KEEP ID1 ID2;
IF ID1=ID2 then delete;
end;
run;
That sort of works, but I still end up with some that end up with duplicates that won't resolve no matter how many times I run (I would have looped it, but I don't know how), because they are just switching back and forth between IDs that have other duplicates (I can get down to about 2,000 of these).
I have figured out that instead of using LAG, I need replace all values after the current line with ID2 for each ID1 value, but I cannot figure out how to do that.
I want to read observation 1, find all later instances of the value of ID1, in both ID1 or ID2 columns, and replace that value with the current observation's ID2 value. Then I want to repeat that process with line 2 and so on.
For the example, I would want to look for any instances after line one of the value 1, and replace it with 2, since that is the surviving ID of that pair - 1 may appear further down multiple times in either of the columns, and I need all them to replaced. Line two would look for later values of 3 and replace them with 4, and so one. The end result should be that an ID number only appears once ever in the ID1 column (though it may appear multiple times in the ID2 column).
ID1 ID2
1 2
3 4
2 5
6 1
1 7
5 8
After first line has been read, data set would look as follows:
ID1 ID2
1 2
3 4
2 5
6 2
2 7
5 8
Reading observation two would make no changes since 3 does not appear again; after observation 3, the set would be:
ID1 ID2
1 2
3 4
2 5
6 5
5 7
5 8
Again, there would be not changes from observation four. but observation 5 would cause the final change:
ID1 ID2
1 2
3 4
2 5
6 5
5 7
7 8
I have tried using the following statement but I can't even tell if I am on the complete wrong track or if I just can't get the syntax figured out.
Data want;
Set have;
Do i=_n_;
ID=ID2;
Replace next var{EUID} where (EUID1=EUID1 AND EUID2=EUID1);
End;
Run;
Thanks for your help!
There is no need to work back and forth thru the data file. You just need to retain the replacement information so that you can process the file in a single pass.
One way to do that is to make a temporary array using the values of the ID variables as the index. That is easy to do for your simple example with small ID values.
So for example if all of the ID values are integers between 1 and 1000 then this step will do the job.
data want ;
set have ;
array xx (1000) _temporary_;
do while (not missing(xx(id1))); id1=xx(id1); end;
do while (not missing(xx(id2))); id2=xx(id2); end;
output;
xx(id1)=id2;
run;
You probably need to add a test to prevent cycles (1 -> 2 -> 1).
For a more general solution you should replace the array with a hash object instead. So something like this:
data want ;
if _n_=1 then do;
declare hash h();
h.definekey('old');
h.definedata('new');
h.definedone();
call missing(new,old);
end;
set have ;
do while (not h.find(key:id1)); id1=new; end;
do while (not h.find(key:id2)); id2=new; end;
output;
h.add(key: id1,data: id2);
drop old new;
run;
Here's an implementation of the algorithm you've suggested, using a modify statement to load and rewrite each row one at a time. It works with your trivial example but with messier data you might get duplicate values in ID1.
data have;
input ID1 ID2 ;
datalines;
1 2
3 4
2 5
6 1
1 7
5 8
;
run;
title "Before making replacements";
proc print data = have;
run;
/*Optional - should improve performance at cost of increased memory usage*/
sasfile have load;
data have;
do i = 1 to nobs;
do j = i to nobs;
modify have point = j nobs = nobs;
/* Make copies of target and replacement value for this pass */
if j = i then do;
id1_ = id1;
id2_ = id2;
end;
else do;
flag = 0; /* Keep track of whether we made a change */
if id1 = id1_ then do;
id1 = id2_;
flag = 1;
end;
if id2 = id1_ then do;
id2 = id2_;
flag = 1;
end;
if flag then replace; /* Only rewrite the row if we made a change */
end;
end;
end;
stop;
run;
sasfile have close;
title "After making replacements";
proc print data = have;
run;
Please bear in mind that as this modifies the dataset in place, interrupting the data step while it is running could result in data loss. Make sure you have a backup first in case you need to roll your changes back.
Seems like this should do the trick and is fairly straight forward. Let me know if it is what you are looking for:
data have;
input id1 id2;
datalines;
1 2
3 4
2 5
6 1
1 7
5 8
;
run;
%macro test();
proc sql noprint;
select count(*) into: cnt
from have;
quit;
%do i = 1 %to &cnt;
proc sql noprint;
select id1,id2 into: id1, :id2
from have
where monotonic() = &i;quit;
data have;
set have;
if (_n_ > input("&i",8.))then do;
if (id1 = input("&id1",8.))then id1 = input("&id2",8.);
if (id2 = input("&id1",8.))then id2 = input("&id2",8.);
end;
run;
%end;
%mend test;
%test();
this might be a little faster:
data have2;
input id1 id2;
datalines;
1 2
3 4
2 5
6 1
1 7
5 8
;
run;
%macro test2();
proc sql noprint;
select count(*) into: cnt
from have2;
quit;
%do i = 1 %to &cnt;
proc sql noprint;
select id1,id2 into: id1, :id2
from have2
where monotonic() = &i;
update have2 set id1 = &id2
where monotonic() > &i
and id1 = &id1;
quit;
proc sql noprint;
update have2 set id2 = &id2
where monotonic() > &i
and id2 = &id1;
quit;
%end;
%mend test2;
%test2();
I have a dataset that has columns like:
a|b|c|d|e
and rows like:
1|3|5|7|9
2|4|6|8|10
How to change it to:
Char|Num|
a|1
a|2
b|3
b|4
c|5
c|6
d|7
d|8
e|9
e|10
Thank you in advance!
You can use PROC TRANSPOSE. The only gotcha is to get what you want you need a BY variable. Easiest thing is to add a record number and use that as your BY.
data have;
input a b c d;
i = _n_;
datalines;
1 2 3 4
5 6 7 8
;
run;
proc transpose data=have out=want(drop=i);
by i;
var a b c d;
run;
how can i perform calculation for the last n observation in a data set
For example if I have 10 observations I would like to create a variable that would sum the last 5 values of another variable. Please do not suggest that I lag 5 times or use module ( N ). I need a bit more elegant solution than that.
with the code below alpha is the data set that i have and bravo is the one i need.
data alpha;
input lima ## ;
cards ;
3 1 4 21 3 3 2 4 2 5
;
run ;
data bravo;
input lima juliet;
cards;
3 .
1 .
4 .
21 .
3 32
3 32
2 33
4 33
2 14
5 16
;
run;
thank you in advance!
You can do this in the data step or using PROC EXPAND from SAS/ETS if available.
For the data step the idea is that you start with a cumulative sum (summ), but keep track of the number of values that were added so far (ninsum). Once that reaches 5, you start outputting the cumulative sum to the target variable (juliet), and from the next step you start subtracting the lagged-5 value to only store the sum of the last five values.
data beta;
set alpha;
retain summ ninsum 0;
summ + lima;
ninsum + 1;
l5 = lag5(lima);
if ninsum = 6 then do;
summ = summ - l5;
ninsum = ninsum - 1;
end;
if ninsum = 5 then do;
juliet = summ;
end;
run;
proc print data=beta;
run;
However there is a procedure that can do all kind of cumulative, moving window, etc calculations: PROC EXPAND, in which this is really just one line. We just tell it to calculate the backward moving sum in a window of width 5 and set the first 4 observations to missing (by default it will expand your series by 0's on the left).
proc expand data=alpha out=gamma;
convert lima = juliet / transformout=(movsum 5 trimleft 4);
run;
proc print data=gamma;
run;
Edit
If you want to do more complicated calculations, you need to carry the previous values in retained variables. I thought you wanted to avoid that, but here it is:
data epsilon;
set alpha;
array lags {5};
retain lags1 - lags5;
/* do whatever calculation is needed */
juliet = 0;
do i=1 to 5;
juliet = juliet + lags{i};
end;
output;
/* shift over lagged values, and add self at the beginning */
do i=5 to 2 by -1;
lags{i} = lags{i-1};
end;
lags{1} = lima;
drop i;
run;
proc print data=epsilon;
run;
I can offer rather ugly solution:
run data step and add increasing number to each group.
run sql step and add column of max(group).
run another data step and check if value from (2)-(1) is less than 5. If so, assign to _num_to_sum_ variable (for example) the value that you want to sum, otherwise leave it blank or assign 0.
and last do a sql step with sum(_num_to_sum_) and group results by grouping variable from (1).
EDIT: I have added a live example of the concept in a bit more compacted way.
input var1 $ var2;
cards;
aaa 3
aaa 5
aaa 7
aaa 1
aaa 11
aaa 8
aaa 6
bbb 3
bbb 2
bbb 4
bbb 6
;
run;
data step1;
set sourcetable;
by var1;
retain obs 0;
if first.var1 then obs = 0;
else obs = obs+1;
if obs >=5 then to_sum = var2;
run;
proc sql;
create table rezults as
select distinct var1, sum(to_sum) as needed_summs
from step1
group by var1;
quit;
In case anyone reads this :)
I solved it the way I needed it to be solved. Although now I am more curious which of the two(the retain and my solution) is more optimal in terms of computing/processing time.
Here is my solution:
data bravo(keep = var1 summ);
set alpha;
do i=_n_ to _n_-4 by -1;
set alpha(rename=var1=var2) point=i;
summ=sum(summ,var2);
end;
run;
Suppose a data are as follows:
A B C
1 3 2
1 4 9
2 6 0
2 7 3
where A B and C are the variable names.
Is there a way to transform the table to
A 1
A 1
A 2
A 2
B 3
B 4
B 6
B 7
C 2
C 9
C 0
C 3
Expanding on the advice from #donPablo, here's how you would code it. Create an array to read across the data, then output each iteration of that array so you end up with the number of rows being the rows * columns from the original dataset. The VNAME function enables you to store the variable name (A, B, C) as a value in a separate variable.
data have;
input A B C;
datalines;
1 3 2
1 4 9
2 6 0
2 7 3
;
run;
data want;
set have;
length var1 $10;
array vars{*} _numeric_;
do i=1 to dim(vars);
var1=vname(vars{i});
var2=vars{i};
keep var1 var2;
output;
end;
run;
proc sort data=want;
by var1;
run;
The least amount of (expensive) development time might be --
Read and store the first row
For each subsequent row
Read the row
Create three records
Until end
Sort
How many times will this be run? Per day/ per year?
What number of rows are there?
Might we save 1 hr / month? 1 min / year? Something will need to read the entire file. Optomize last. Make it work first.
tkx
It should work correctly:
DATA A(keep A);
new_var = 'A';
SET your_data;
RUN;
DATA B(keep B);
new_var = 'B';
SET your_data;
RUN;
DATA C(keep C);
new_var = 'C';
SET your_data;
RUN;
PROC APPEND base=A data=B FORCE;
RUN;
PROC APPEND base=A data=C FORCE;
RUN;
Data A is a result data set.