%let list_2=‘a’,’b’,’c’;
PROC SQL;
CREATE TABLE test AS
SELECT
Bin, /*value is a*/
Rank, /*value is 2*/
CASE
WHEN Bin IN (&list_Rank.) /*list_2=‘a’,’b’,’c’*/
THEN 1
ELSE 0
END AS test
FROM Source
;
QUIT;
I am looking for a way to use the value of column Rank in the reference to the macrovariable list_2.
As #John commented, you can use SYMGET function to use data to look up the value of a macro variable during run-time. I don't think you can use the IN operator for what you want, because it expects a list of character strings, while SYMGET will return a single string. Below I took the quotes out of the &LIST_2 (just to make it easier to work with), and used FINDW to do the work of IN. I think this is along the lines of what you're describing:
%let list_2=a,b,c;
%let list_3=d,e,f ;
data have ;
input bin $1. rank ;
cards ;
a 2
a 3
e 2
e 3
;
run ;
proc sql ;
select
bin
,rank
,case when findw(symget(cats("list_",rank))
,bin
) then 1
else 0
end as test
,symgetc(cats("list_",rank)) as symgetc /*debugging*/
from have ;
quit ;
You could add another CASE statement. Actually since it is SAS code you could eliminate the current CASE statement and just use the result of the IN operation which will be 0 (false) or 1 (true).
%let list_1='a','b','c';
%let list_2='d','e','f';
%let nlists=2;
PROC SQL;
CREATE TABLE test AS
SELECT
Bin /*value is a*/
, Rank /*value is 2*/
, case
when (rank=1) then bin in (&list_1)
when (rank=2) then bin in (&list_2)
else 0
end as test
FROM Source
;
QUIT;
If you are running inside a macro you could use a %do loop to generate the when clauses.
%do i=1 %to &nlists ;
when (rank=&i) then bin in (&&list_&i)
%end;
If you are not in a macro then create a macro to run the %do loop and call it at the appropriate place.
Related
I have a SAS Script containing multiple PROC SQLs. The question is that the SQL Query should be "adapted" based on a SAS variable value, for example :
%let COD_COC =(52624, 52568, 52572);
%let COD_BLOC = ();
proc sql;
create table work.abordados as
select t1.cd_acao,
t1.cd_bloc,
t1.cd_tip_cnl,
t1.cd_cmco,
t1.cd_cli,
datepart(t1.ts_ctt) format=date9. as data_abordagem,
intnx('day',datepart(t1.ts_ctt), &Prazo_Compra) format=date9. as data_limite
from db2coc.ctt_cli_cmco t1
where (t1.cd_acao in &COD_COC)
and (t1.cd_bloc in &COD_BLOC) <<<<<<< facultative filter
;quit;
The question is that the second filter (t1.cd_bloc in &COD_BLOC) should be applied only if the %let COD_BLOC = (); is different of "()".
I´ve been reading about "match/case" on SQL but as far as I know, this test applies to results of queries/values. On my case, what I must test is the SAS variable.
How handle this?
Knowing you want to apply the COD_BLOC in-list filter only when there are one or more values, AND that a proper in-list will have at least 3 source code characters (*), you can test the length as the criteria for using the macro variable.
When the %IF is in open code you need a %do %end block as follows:
...
%if %length(&COD_BLOC) > 2 %then %do;
and t1.cd in &COD_BLOC
%end;
...
When the code is inside a macro, you can use the above or the below
...
%if %length(&COD_BLOC) > 2 %then
and t1.cd in &COD_BLOC
;
...
Another possible coding solution is to use %sysfunc(IFC(...)) to conditionally generate code
...
%sysfunc(ifc(%length(&COD_BLOC) > 2
, and t1.cd in &COD_BLOC
, %str()
))
...
Two good ways to do this, I think.
First: the easy hack.
%let COD_COC =(52624, 52568, 52572);
%let COD_BLOC = (and t1.cd_bloc in (...));
%let COD_BLOC = ;
proc sql;
create table work.abordados as
select t1.cd_acao,
t1.cd_bloc,
t1.cd_tip_cnl,
t1.cd_cmco,
t1.cd_cli,
datepart(t1.ts_ctt) format=date9. as data_abordagem,
intnx('day',datepart(t1.ts_ctt), &Prazo_Compra) format=date9. as data_limite
from db2coc.ctt_cli_cmco t1
where (t1.cd_acao in &COD_COC)
&COD_BLOC <<<<<<< facultative filter
;quit;
Tada, now it is just ignored. Comment out or delete the second line if you want it to be used (and put values in the ... ).
Second, the more proper way, is to use the macro language. This is more commonly done in a macro, but in 9.4m7 (the newest release, and a few years old now) you can do this in open code.
%let COD_COC =(52624, 52568, 52572);
%let COD_BLOC = ();
proc sql;
create table work.abordados as
select t1.cd_acao,
t1.cd_bloc,
t1.cd_tip_cnl,
t1.cd_cmco,
t1.cd_cli,
datepart(t1.ts_ctt) format=date9. as data_abordagem,
intnx('day',datepart(t1.ts_ctt), &Prazo_Compra) format=date9. as data_limite
from db2coc.ctt_cli_cmco t1
where (t1.cd_acao in &COD_COC)
%if %sysevalf(%superq(COD_BLOC) ne %nrstr(%(%)),boolean) %then %do;
and (t1.cd_bloc in &COD_BLOC) <<<<<<< facultative filter
%end;
;quit;
You have to be careful with the ne () bit because () are macro language syntax elements, hence the long %nrstr to make sure they're properly considered characters. (%str would be okay too, I just default to %nrstr.)
I need to select the name of states that do not start with M if the the macro variable M=N but return only the names of states that start with M if macro variable is equal to any other variable using conditional processing.
for example:
%let M=N;
proc sql;
select states,profit,
case
when .....
else
end
from geography_dim
quit;
For the sake of argument, suppose you change the name of the macro variable M to something more ridiculously expressive, such as YN_OPTION_SELECT_M_STATES
%let YN_OPTION_SELECT_M_STATES = N;
proc sql;
select states,profit,
case
when .....
else
end
from geography_dim
/* add this */
where
("&YN_OPTION_SELECT_M_STATES" eq 'N' & STATE not like 'M%')
or
("&YN_OPTION_SELECT_M_STATES" ne 'N' & STATE like 'M%')
;
quit;
Revert to macro variable M if you must, however the code will be somewhat opaque.
It is not SQL but very simple in datastep. If you want check the staring with M macro values in that case "N" you can do like:
/*test data*/
data geography_dim ;
states="Aaaaa";profit=10;output;
states="Naaaa";profit=10;output;
run;
/*set macro variable*/
%let M=N;
/*check if you want*/
%put "&M";
/*your case in datastep*/
data test;
set geography_dim;
if substr(states,1,1) eq "&M" then profit=profit*10;
else profit=0;
run;
/* results
states profit
Aaaaa 0
Naaaa 100
*/
I have a data set named test with variables region, X1, X2, X3, X4......X20.
I would like to take the aggregates of X1-X20 by region level. My code is as follow:
proc sql;
create table test2 as
select sum(X1) as X1, sum(X2) as X2,......, sum(X20) as X20
from test
group by region;
quit;
Would there a faster way to get the sum as opposed to manually typing the sum for 20 of each variables? Thank you!
Why use PROC SQL instead of PROC MEANS?
proc means noprint nway data=test ;
class region ;
var x1-x20;
output out=test2 sum=;
run;
A good use case for SAS macros. Here's a link to a tutorial, but I also highly recommend the purple Carpenter book.
%MACRO do_test2;
proc sql;
create table test2 as
select
%DO i = 1 %TO 20;
sum(X&i) as X&1
%IF &i ^= 20 %THEN , ;
%END;
from test
group by region;
quit;
%MEND;
As #user667489 points out, because the macro code is using a %DO loop, the code block must be defined as a named, invokable macro (%MACRO do_test2 . . . %MEND). Once you've defined it as above, you call it like this:
%do_test2;
Since we have this requirement, we can take advantage of the situation and make the number of repetitions a parameter, in case you ever want to vary it. Then the definition is:
%MACRO do_test2(num_reps);
proc sql;
create table test2 as
select
%DO i = 1 %TO &num_reps;
sum(X&i) as X&1
%IF &i ^= &num_reps %THEN , ;
%END;
from test
group by region;
quit;
%MEND;
Which you would call like this:
%do_test2(20);
So, i want to have a macro that has others macros inside.
Here is the code: `
proc sql NOPRINT ;
select id into :l_id separated by ' ' from work.AMOSTRACHU;
select count(*) into :nr_reg separated by ' ' from tdata.work.AMOSTRACHU;
quit;
* check;
%put l_id=&l_id nr_reg=&nr_reg;
%macro ciclo_first();
%do n=1 %to &nr_reg;
%let ref=%scan(&l_id,&n);
%put ref=&ref;
proc sql;
select recetor into : lsus&ref separated by ' ' from tdata.5pct_&ref;
select count(*) into :nrsus&ref separated by ' ' from tdata.5pct_&ref;
quit;
%put lsus&ref=&lsus&ref;
%put nrsus&ref=&nrsus&ref;
%MACRO CICLO_PF_SUSref();
%do n=1 %to &nrsus&ref %by 1;
%let sus=%scan(&lsus&ref,&n);
%put sus=&sus;
%LET I = %EVAL(14);
%DO %WHILE (&I<=24);
*my code (depends on &i and &sus)* (works fine alone)
%LET I = %EVAL(&I+1);
%END;
%END;
%MEND;
%CICLO_PF_SUSref;
%MACRO CICLO_PF_SUS_CSRANK();
%do n=1 %to &nrsus&refm %by 1;
%let sus=%scan(&lsus&ref,&n);
%put sus=&sus;
%CICLO_PF_SUSPEITOSrefmsisdn;
%CICLO_PF_SUS_CSRANK;
my code ( just depends on &sus)/
%END;
%MEND;
%CICLO_PF_SUS_CSRANK;
%end;
%mend;
%ciclo_first;`
I think the major problem is in this part:
%put lsus&ref=&lsus&ref;
%put nrsus&ref=&nrsus&ref;
And the error about that is:
A character operand was found in the %EVAL function or %IF condition
where a numeric operand is required. The condition was:
&nrsus&ref
How can i change this in order to work? I understand that it doesn't make all the sense to have something depending on two, like &nrsus&ref.
the first warnings and errors appears here:
ref=15
WARNING: Apparent symbolic reference LSUS not resolved.
lsus15=&lsus15 WARNING: Apparent symbolic
reference NRSUS not resolved.
nrsus15=&nrsus15 ERROR: Expected semicolon not
found. The macro will not be compiled.
How can i solve this? Have no ideas and it would be really useful to make this macro functional in order to avoid to run this 100 times.
UPDATE [06.08.2015]
I have a table with 100 numbers, that's in
'work.amostrachu'.
I created the macro ciclo_first in order to run the other 2 macros for this list. because, if i replace manually the &ref by the number i want it works fine.
Let's suppose 'work.amostrachu' has:
ID 1 2 3 (...) till n=100
Then, with this part:
proc sql;
select recetor into : lsus&ref separated by ' ' from work.5pct_&ref;
select count(*) into :nrsus&ref separated by ' ' from work.5pct_&ref;
quit;
I want to get the elements that are on the column 'recetor' of work.5pct_&ref.
For ID=1 i would obtain lsus1 composed by, for example, 3 numbers (124,564,859)
And, then, the %MACRO CICLO_PF_SUSref(); will have as input these 3 numbers (that could be 4 or 5 or sometingh else).
(here, i might be calling badly the list of elements i want from 'work.5pct_&ref).
Then, the output of the previous macro would be the input of this one: %MACRO CICLO_PF_SUS_CSRANK.
And that would be all.
The %MACRO CICLO_PF_SUSref() and %MACRO CICLO_PF_SUS_CSRANK works ok if i just replace the &ref by the id. that's why i tried to create a macro that would run these 2 macros for the initial list. if you have best ideas, i would be thankful.
So, i want something that allows me to run this two macros (%MACRO CICLO_PF_SUSref() and `%MACRO CICLO_PF_SUS_CSRANK) for the list i get in the beginning:
proc sql NOPRINT ;
select id into :l_id separated by ' ' from work.AMOSTRACHU;
select count(*) into :nr_reg separated by ' ' from tdata.work.AMOSTRACHU;
quit;
[UPDATE 10.08.2015]
Ok, just read the suggested answers and worked on it.
I have a list, with the identification(numerical) of 100 clients, let's call each client : ref. That's on WORK.AMOSTRACHU.
I wroted the following code and it worked, and will help me explain you what i want:
proc sql NOPRINT ;
select id into :l_id separated by ' ' from work.AMOSTRACHU;
select count(*) into :nr_reg separated by ' ' from work.AMOSTRACHU;
quit;
* check;
%put l_id=&l_id nr_reg=&nr_reg;
%macro lista_ent();
%do n=1 %to &nr_reg;
%put n=&n;
%let ref=%scan(&l_id,&n);
%put ref=&ref;
proc sql;
select recetor into :listae&ref SEPARATED BY ' ' from work.e5pct_id&ref;
select count(*) into :nre&ref separated by ' ' from work.e5pct_id&ref;
quit;
%end;
%mend;
%lista_ent;
Will show you the output for the first 3 cases (of 100, the beggining list in work.amostrachu), it's the results part in SAS:
Recetor
507
723
955
-page break-
3
-page break-
380
500
675
977
984
-page break-
5
-page break-
200
225
351
488
698
781
927
-page break-
7
So, i have the 'values' of the column 'recetor' of the data work.e5pct_id&ref and how many values i have for each ref. (i've showed you results for the first 3 refs, but i have it for the 100).
Now, the first macro:
%MACRO CICLO_M_PF_ref();
%local me n i;
%do n=1 %to nre&ref %by 1;
%let me=%scan(listae&ref,&n);
%put me=&me;
%LET I = %EVAL(14);
%DO %WHILE (&I<=24);
proc sql;
create table work.smthng_&I as
select * from
work.wtv&I
WHERE A=&me OR B=&me;RUN;
PROC APPEND
DATA=work.smthng_&I
BASE=work.pf_&me
FORCE;
RUN;
%LET I = %EVAL(&I+1);
%END;
%END;
%MEND;
%CICLO_M_PF_ref;
My all doubts in the & and && are around here.
So, with the data: I have my first ref whose results of column 'recetor' are
Recetor
507
723
955
-page break-
3
So, i want to run that code for each one of this values. First for '507', then for '723' and then for '955', and i want to do it for all the refs.
So, when the macro finishes to run my code for this 3, i want the macro to skip to the second ref and then run my code for the values of the column 'recetor' for the second ref: 380,500,675,977 and 984.
i used this code:
proc sql;
select recetor into :listae&ref SEPARATED BY ' ' from work.e5pct_id&ref;
select count(*) into :nre&ref separated by ' ' from work.e5pct_id&ref;
quit;
because each one of the refs have different values and the number of them could be different, just as i showed you. so, this whas to tell the macro to run it nre&ref times and for all values in the list listae&ref.
the error is the following:
ERROR: A character operand was found in the %EVAL function or %IF
condition where a numeric operand is required. The condition was:
nre&ref ERROR: The %TO value of the %DO T loop is invalid. ERROR: The macro CICLO_M_PF_REF will stop executing.
I can't quite follow your desired output and macro but here are some things I noticed.
None of your macros take parameters. If you change your macro to take parameters you can call them individually which may help to stream line your process.
I think you want something like this:
%macro def1(param1);
...
%mend;
%macro def2(param2);
...
%mend;
%macro execute();
%do i=1 to 100;
%def1(param1);
%def2(param2);
%end;
%mend;
This still seems a bit awkward, so if you can explain your process with your data there may be a better way overall.
I see a number of issues you could address, but without test data it is hard to evaluate.
When trying to show the value for macro variable x&i you need to double up on the prefix &. So if I=1 and X1 = FRED then &&x&i = FRED.
When pushing values into macro variables from SQL use the automatic macro variable SQLOBS to get the record count. No need to run the query again to get the count.
You cannot select COUNT(*) into multiple macro variables. SQL will just return one count.
SAS dataset or variable names cannot start with a digit (tdata.5pct_&ref) or contain periods (tdata.work.AMOSTRACHU).
Do NOT nest macro definitions. You can nest the calls, but nesting
the definitions is just going to lead to confusion.
Your actual nested macros do not make much sense. What is this variable I that is introduced? It appears to be a constant.
Why not just code them as part of the outer macro? Not much need to make them separate macros if they are only called at one place.
If you do nest them then make sure to define your local macro variables as local to prevent overwriting the values of macro variables with the same name that might exist in an outer macro scope. The N looping variable for your %DO loops for example.
First define your subroutine macros.
%MACRO CICLO_PF_SUSref(ref_list);
* CICLO_PF_SUSref ;
%local n sus;
%do n=1 %to %sysfunc(countw(&ref_list,%str( )));
%let sus=%scan(&ref_list,&n);
%put NOTE: &sysmacroname N=&n SUS=&sus;
%end;
%MEND CICLO_PF_SUSref;
%MACRO CICLO_PF_SUS_CSRANK(ref_list);
* CICLO_PF_SUS_CSRANK ;
%local n sus ;
%do n=1 %to %sysfunc(countw(&ref_list,%str( )));
%let sus=%scan(&ref_list,&n);
%put NOTE: &sysmacroname N=&n SUS=&sus;
%put NOTE: Call macro named: CICLO_PF_SUSPEITOSrefmsisdn;
%end;
%MEND CICLO_PF_SUS_CSRANK;
Then your main macro.
%macro ciclo_first(id_list);
* Start ciclo_first ;
%local n id ;
%do n=1 %to %sysfunc(countw(&id_list,%str( )));
%let id=%scan(&id_list,&n);
proc sql noprint;
select recetor into : lsus&id separated by ' ' from pct_&id;
%let nrsus&id = &sqlobs ;
quit;
%put NOTE: Current ID=&id ;
%put NOTE: &&nrsus&id records read from PCT_&ID ;
%put NOTE: Value List LSUS&id = &&LSUS&id ;
%CICLO_PF_SUSref(&&lsus&id);
%CICLO_PF_SUS_CSRANK(&&lsus&id);
%end;
* End ciclo_first ;
%mend ciclo_first;
Then setup some data and call the main macro.
* Setup test data ;
data AMOSTRACHU;
do id=1 to 2; output; end;
run;
data PCT_1 ;
do recetor='A','B';
output;
end;
run;
data PCT_2 ;
do recetor='C','D';
output;
end;
run;
options mprint;
%ciclo_first(1 2);
I need to get a percentage for 75 values in 75 columns individually. And I want to use a do loop so I don't have to hard code it 75 times. There are some conditions so there will be a where statement.
I am not getting the do loop correctly but I am using the below to get a percentage
case when (SUM(t1.sam)) >0 then
((SUM(t1.sam))/(SUM(t1.sam_Threshold)))*100
else 0
end
I tried the below and its a bit better:
data test;
i_1=4;
i_2=8;
i_3=4;
i_4=8;
V_ANN_V_INSP=24;
run;
%macro loop();
%let numcols=4;
proc sql;
create table test3 as
select V_ANN_V_INSP,
%do i=1 %to &numcols;
(i_&i/V_ANN_V_INSP)*100 as i_&i._perc
%if &i<&numcols %then %do;,
%end;
%end;
from test;
quit;
%mend;
%loop();
CASE WHEN is a SQL statement, not a data step statement, so you can't use a DO loop there. Depending on what you're doing exactly, there are a lot of possible solutions here. Posting additional code would help to get a more precise answer, but I can give you a few suggestions.
First, take it into a data step. Then you can use a do loop.
data want;
set have;
array nums sam1-sam75;
array denoms threshold1-threshold75;
array pct[75];
do _t = 1 to dim(nums);
pct[_t]=nums[_t]/denoms[_t];
end;
run;
Second, if you need to do this in SQL for some reason, you can write out the SQL code either in a macro or in a data step in a pre-processing step.
%macro do_sql_st;
%do _t = 1 to 75;
case when (SUM(t1.sam&_t.)) >0 then
((SUM(t1.sam&_t.))/(SUM(t1.sam_Threshold&_t.)))*100
else 0
end
as pct&_t.
%end;
%mend do_sql_st;
proc sql;
select %do_sql_st from t1 where ... ;
quit;
These are not terribly flexible; unless you have very specifically named variables, they won't work as is. You're more likely to want to do some sort of data step preprocessing I suspect, but that's very hard to explain without more detail as to how the variables are named (ie, if there is a relationship between them).