I have a bunch of character variables which I need to sort out from a large dataset. The unwanted variables all have entries that are the same or are all missing (meaning I want to drop these from the dataset before processing the data further). The data sets are very large so this cannot be done manually, and I will be doing it a lot of times so I am trying to create a macro which will do just this. I have created a list macro variable with all character variables using the following code (The data for my part is different but I use the same sort of code):
data test;
input Obs ID Age;
datalines;
1 2 3
2 2 1
3 2 2
4 3 1
5 3 2
6 3 3
7 4 1
8 4 2
run;
proc contents
data = test
noprint
out = test_info(keep=name);
run;
proc sql noprint;
select name into : testvarlist separated by ' ' from test_info;
quit;
My idea is then to just use a data step to drop this list of variables from the original dataset. Now, the problem is that I need to loop over each variable, and determine if the observations for that variable are all the same or not. My idea is to create a macro that loops over all variables, and for each variable counts the occurrences of the entries. Since the length of this table is equal to the number of unique entries I know that the variable should be dropped if the table is of length 1. My attempt so far is the following code:
%macro ListScanner (org_list);
%local i next_name name_list;
%let name_list = &org_list;
%let i=1;
%do %while (%scan(&name_list, &i) ne );
%let next_name = %scan(&name_list, &i);
%put &next_name;
proc sql;
create table char_occurrences as
select &next_name, count(*) as numberofoccurrences
from &name_list group by &next_name;
select count(*) as countrec from char_occurrences;
quit;
%if countrec = 1 %then %do;
proc sql;
delete &next_name from &org_list;
quit;
%end;
%let i = %eval(&i + 1);
%end;
%mend;
%ListScanner(org_list = &testvarlist);
Though I get syntax errors, and with my real data I get other kinds of problems with not being able to read the data correctly but I am taking one step at a time. I am thinking that I might overcomplicate things so if anyone has an easier solution or can see what might be wrong to I would be very grateful.
There are many ways to do this posted around.
But let's just look at the issues you are having.
First for looping through your space delimited list of names it is easier to let the %do loop increment the index variable for you. Use the countw() function to find the upper bound.
%do i=1 %to %sysfunc(countw(&name_list,%str( )));
%let next_name = %scan(&name_list,&i,%str( ));
...
%end;
Second where is your input dataset in your SQL code? Add another parameter to your macro definition. Where to you want to write the dataset without the empty columns? So perhaps another parameter.
%macro ListScanner (dsname , out, name_list);
%local i next_name sep drop_list ;
Third you can use a single query to count all of variables at once. Just use count( distinct xxxx ) instead of group by.
proc sql noprint;
create table counts as
select
%let sep=;
%do i=1 %to %sysfunc(countw(&name_list,%str( )));
%let next_name = %scan(&name_list,&i,%str( ));
&sep. count(distinct &next_name) as &next_name
%let sep=,;
%end;
from &dsname
;
quit;
So this will get a dataset with one observation. You can use PROC TRANSPOSE to turn it into one observation per variable instead.
proc transpose data=counts out=counts_tall ;
var _all_;
run;
Now you can just query that table to find the names of the columns with 0 non-missing values.
proc sql noprint ;
select _name_ into :drop_list separated by ' '
from counts_tall
where col1=0
;
quit;
Now you can use the new DROP_LIST macro variable.
data &out ;
set &dsname ;
drop &drop_list;
run;
So now all that is left is to clean up after your self.
proc delete data=counts counts_tall ;
run;
%mend;
As far as your specific initial question, this is fairly straightforward. Assuming &testvarlist is your macro variable containing the variables you are interested in, and creating some test data in have:
%let testvarlist=x y z;
data have;
call streaminit(7);
do id = 1 to 1e6;
x = floor(rand('Uniform')*10);
y = floor(rand('Uniform')*10);
z = floor(rand('Uniform')*10);
if x=0 and y=4 and z=7 then call missing(of x y z);
output;
end;
run;
data want fordel;
set have;
if min(of &testvarlist.) = max(of &testvarlist.)
and (cmiss(of &testvarlist.)=0 or missing(min(of &testvarlist.)))
then output fordel;
else output want;
run;
This isn't particularly inefficient, but there are certainly better ways to do this, as referenced in comments.
Related
Can anyone help with this issue I'm having where the macro is only taking the final row value of the data?
I have some data that looks like this:
data data1 ;
infile datalines dsd dlm='|' truncover;
input id :$2. year_age_15 EDU_2000 EDU_2001 EDU_2002 ;
datalines4;
10|2000|3|4|5
11|2000|5|5|6
12|2001|1|2|3
13|2002|5|5|6
14|2001|2|2|2
15|2000|3|3|4
;;;;
However I need it to use the year variable to determine which data to keep, and then change all the values for the years after that value to missing, like so:
data data1 ;
infile datalines dsd dlm='|' truncover;
input id :$2. year_age_15 EDU_2000 EDU_2001 EDU_2002 ;
datalines4;
10|2000|3|.|.
11|2000|5|.|.
12|2001|1|2|.
13|2002|5|5|6
14|2001|2|2|.
15|2000|3|.|.
;;;;
I've been trying to get this macro to work, but it only works intermittently and works just for the final row of the data rather than looping through the rows.
%macro macro2 (output=, input=);
data &output;
set &input;
%DO I = 1 %TO 6;
%do; call symput('value2',trim(left(put(year_age_15,8.))));
temp_col=&value2.;
%let year_end=&value2.;
%put YEAR END IS: &year_end.;
%put EDU YEAR IS: EDU_&year_end.;
%do year = &year_end. %TO 2002;
%put &year.;
EDU_&year.=.;
%end;
%end;
%end;
run;
%MEND macro2;
%macro1(input=testset, output=output_testset);
In R it could be something simple like :
for(i in 1:6){.
do this
}
Any advice? I can't figure out which bit is going wrong, thanks!
So, I think the issue here is your data is at the wrong level. You certainly can do what Reeza suggests, and I think it's probably reasonable to do so, but the reason why this is a bit complicated is that you have data in your variable name. That's not a best practice - your variable name should be "education" and your data should have a row for each year. Then this would be a simple WHERE statement!
Here's a simple PROC TRANSPOSE that turns it to the right structure, and then if you really need it the other way, a second one will turn it back. The where statement can be in the proc transpose or could be used somewhere else.
proc transpose data=data1 out=data_t (where=(year_Age_15 ge input(scan(_NAME_,2,'_'),4.)));
by id year_Age_15;
var edu_:;
run;
proc transpose data=data_t out=want;
by id year_age_15;
id _name_;
var col1;
run;
Create an array and index it by years rather than default 1:n
Loop through your array starting at year+1 and set to missing
data want;
set data1;
array educ(2000:2002) edu_2000-edu_2002;
if (year_age_15 +1) <= hbound(educ) then do i= (year_age_15 +1) to hbound(educ);
call missing(educ(i));
end;
run;
As #Joe mentions, the year to match is part of a variable name, which is tremor inducing 'data in the metadata'
You can use the VNAME to retrieve the variable name of an index accessed array element. Use that feature to compare to expected variable name whilst looping over a variable array based on variables named EDU*.
Example:
data have ;
infile datalines dsd dlm='|' truncover;
input id :$2. year_age_15 EDU_2000 EDU_2001 EDU_2002 ;
datalines4;
10|2000|3|4|5
11|2000|5|5|6
12|2001|1|2|3
13|2002|5|5|6
14|2001|2|2|2
15|2000|3|3|4
;;;;
data want;
set have;
array edus edu_:;
* find index of element corresponding to variable name having year;
do _n_ = 1 to dim(edus) until (upcase(vname(edus(_n_))) = cats('EDU_',year_age_15));
end;
* fill in elements at indices post the found one with missing values;
do _n_ = _n_+1 to dim(edus);
call missing(edus(_n_));
end;
run;
So I have a dataset with one primary key: unique_id and 1200 variables. This dataset is generated from a macro so the number of columns will not be fixed. I need to split this dataset into 4 or more datasets of 250 variables each, and each of these smaller datasets should contain the primary key so that I can merge them back later. Can somebody help me with either a sas function or a macro to solve this?
Thanks in advance.
A simple way to split a datasets in the way you request is to use a single data step with multiple output datasets where each one has a KEEP= dataset option listing the variables to keep. For example:
data split1(keep=Name Age Height) split2(keep=Name Sex Weight);
set sashelp.class;
run;
So you need to get the list of variables and group then into sets of 250 or less. Then you can use those groupings to generate code like above. Here is one method using PROC CONTENTS to get the list of variables and CALL EXECUTE() to generate the code.
I will use macro variables to hold the name of the input dataset, the key variable that needs to be kept on each dataset and maximum number of variables to keep in each dataset.
So for the example above those macro variable values would be:
%let ds=sashelp.class;
%let key=name;
%let nvars=2;
So use PROC CONTENTS to get the list of variable names:
proc contents data=&ds noprint out=contents; run;
Now run a data step to split them into groups and generate a member name to use for the new split dataset. Make sure not to include the KEY variable in the list of variables when counting.
data groups;
length group 8 memname $41 varnum 8 name $32 ;
group +1;
memname=cats('split',group);
do varnum=1 to &nvars while (not eof);
set contents(keep=name where=(upcase(name) ne %upcase("&key"))) end=eof;
output;
end;
run;
Now you can use that dataset to drive the generation of the code:
data _null_;
set groups end=eof;
by group;
if _n_=1 then call execute('data ');
if first.group then call execute(cats(memname,'(keep=&key'));
call execute(' '||trim(name));
if last.group then call execute(') ');
if eof then call execute(';set &ds;run;');
run;
Here are results from the SAS log:
NOTE: CALL EXECUTE generated line.
1 + data
2 + split1(keep=name
3 + Age
4 + Height
5 + )
6 + split2(keep=name
7 + Sex
8 + Weight
9 + )
10 + ;set sashelp.class;run;
NOTE: There were 19 observations read from the data set SASHELP.CLASS.
NOTE: The data set WORK.SPLIT1 has 19 observations and 3 variables.
NOTE: The data set WORK.SPLIT2 has 19 observations and 3 variables.
Just another way of doing it using macro variables:
/* Number of columns you want in each chunk */
%let vars_per_part = 250;
/* Get all the column names into a dataset */
proc contents data = have out=cols noprint;
run;
%macro split(part);
/* Split the columns into 250 chunks for each part and put it into a macro variable */
%let fobs = %eval((&part - 1)* &vars_per_part + 1);
%let obs = %eval(&part * &vars_per_part);
proc sql noprint;
select name into :cols separated by " " from cols (firstobs = &fobs obs = &obs) where name ~= "uniq_id";
quit;
/* Chunk up the data only keeping those varaibles and the uniq_id */
data want_part∂
set have (keep = &cols uniq_id);
run;
%mend;
/* Run this from 1 to whatever the increment required to cover all the columnns */
%split(1);
%split(2);
%split(3);
this is not a complete solution but some help to give you another insight into how to solve this. The previous solutions have relied much on proc contents and data step, but I would solve this using proc sql and dictionary.columns. And I would create a macro that would split the original file into as many parts as needed, 250 cols each. The steps roughly:
proc sql; create table as _colstemp as select * from dictionary.columns where library='your library' and memname = 'your table' and name ne 'your primary key'; quit;
Count the number of files needed somewhere along:
proc sql;
select ceil(count(*)/249) into :num_of_datasets from _colstemp;
select count(*) into :num_of_cols from _colstemp;
quit;
Then just loop over the original dataset like:
%do &_i = 1 %to &num_of_datasets
proc sql;
select name into :vars separated by ','
from _colstemp(firstobs=%eval((&_i. - 1)*249 + 1) obs = %eval(min(249,&num_of_cols. - &_i. * 249)) ;
quit;
proc sql;
create table split_&_i. as
select YOUR_PRIMARY_KEY, &vars from YOUR_ORIGINAL_TABLE;
quit;
%end;
Hopefully this gives you another idea. The solution is not tested, and may contain some pseudocode elements as it's written from my memory of doing things. Also this is void of macro declaration and much of parametrization one could do.. This would make the solution more general (parametrize your number of variables for each dataset, your primary key name, and your dataset names for example.
First of all, apologize for my poor english but that's because I'm not native. I'm a newbie in SAS programming too, and I need someone to help me with this problem struggling me.
I have one dataset A containing a numeric field YM representing year and month (e.g., 200902) that I'm using to filter the dataset. In particular, I want to get N filtered datasets using N differents values YM.
A_filtered_200901 = A.filter(YM == 200901)
A_filtered_200902 = A.filter(YM == 200902)
A_filtered_200903 = A.filter(YM == 200903)
...
My idea was to generate the sequence of YM used to filter and then give it as an argument to a %macro containing a PROC SQL. In code/pseucode:
data ym_dataset;
date = input(put(20090201, 8.), yymmdd8.);
do i = 1 to 3;
aux1 = intnx('MONTH', date, i);
aux2 = put(aux1, yymmddn8.);
list_of_ym_values = substr(aux2 , 1, 6);
output;
end;
run;
%macro my_macro(list_of_ym_values);
proc sql;
%do i = 1 %to dim(&list_of_ym_values)
select *
from A
where YM = &list_of_ym_values(i)
%end
quit;
%mend my_macro;
%my_macro(ym_dataset[list_of_ym_values])
I know that this is not the correct approach, but I hope that someone could shed me some light about doing it properly.
Thank you!!
you need loop through list of variables and this values can be created in a macro variable. But as #richard suggested in comments is not great idea to split datasets.
/* create macrovariable with all values*/
proc sql;
select list_of_ym_values into :List
separated by "|" from ym_dataset;
%put &list;
/* scan through each variable and create new dataset*/
%macro one;
%do i=1 %to %sysfunc(countw(&list),"|") ;
%let val= %scan(&list,&i,|);
proc sql;
create table want_&val as
select * from ym_dataset
where list_of_ym_values = "&val";
%end;
%mend;
%one;
Is there any form to keep variables with a doop loop in data step?
will be something as:
data test;
input id aper_f_201501 aper_f_201502 aper_f_201503 aper_f_201504
aper_f_201505 aper_f_201506;
datalines;
1 0 1 2 3 5 7
2 -1 5 4 8 7 9
;
run;
%macro test;
%let date = '01Jul2015'd;
data test2;
set test(keep=do i = 1 to 3;
aper_f_%sysfunc(intnx(month,&date,-i,begin),yymmn6.);
end;)
run;
%mend;
%test;
I need to iterate several dates.
Thank you very much.
You need to use macro %do loop instead of the data step do loop, which is not going to be valid in the middle of a dataset option. Also do not generate those extra semi-colons into the middle of your dataset options. And do include a semi-colon to end your SET statement.
%macro test;
%local i date;
%let date = '01Jul2015'd;
data test2;
set test(keep=
%do i = 1 %to 3;
aper_f_%sysfunc(intnx(month,&date,-i,begin),yymmn6.)
%end;
);
run;
%mend;
%test;
You can use the colon shortcut to reference variables with the same prefix, anything in front of the colon will be kept.
keep ID aper_f_2015: ;
There's also a hyphen when you have sequential lists
keep ID aper_f_201501-aper_f_201512;
You can use a macro but not sure it adds a lot of value here.
I'm trying to modify a number of datasets (their names follow a certain order, like data_AXPM061203900_20120104 , data_AXPM061203900_20120105, data_AXPA061204100_20120103, data_AXPA061204100_20120104) under work library. For example, I want to delete variable "price=0" in all datasets.
I am using the following to create a table to identify the datasets:
proc sql ;
create table data.mytables as
select *
from dictionary.tables
where libname = 'WORK'
order by memname ;
quit ;
For the next step, I'm trying to use a macro:
%macro test;
proc sql ;
select count(memname) into: obs from data.mytables;
%let obs=&obs.;
select catx("_", "data", substr(memname, 6, 13), substr(memname,20,27))
into :setname1-:setname&obs.
from data.mytables;
quit;
%do i=1 %to &obs.;
data &&setname&i
set &&setname&i
if bid_price= '.' then delete;
%end;
%mend test;
However, it completely failed. Could anyone give me some suggestions? I'm really not good at macros. Errors include:
56: LINE and COLUMN cannot be determined.
ERROR 56-185: SET is not allowed in the DATA statement when option DATASTMTCHK=COREKEYWORDS. Check for a missing semicolon in
the DATA statement, or use DATASTMTCHK=NONE.
You are missing semicolons on the DATA statement and SET statement, and should probably add a RUN statement. Suggest you try:
%do i=1 %to &obs.;
data &&setname&i ;
set &&setname&i ;
if bid_price= '.' then delete;
run;
%end;
Note that the DELETE statement is deleting records, not deleting variables. The code above expects that bid_price is a character variable, and that you want to delete records when the value is '.'.