DILIP KUMAR
DILIP KUMAR

Reputation: 45

FIND LOGIC FOR SORTING DATAS

I am a beginner in SAS and I have some datelines as

VIT_A 12 23 22 0 32 0 11
VIT_C 10 9 0 21 0 26 76

a

if I want to list only the datas where VIT_A ^=0 , VIT_C^=0 what will be the logic?

Results will be

VIT_A 12 23 22 32 11
VIT_C 10 9 21 26 76

Upvotes: 0

Views: 54

Answers (3)

momo1644
momo1644

Reputation: 1804

If your data looks like this Table 1:

    COL1    COL2
1   VIT_A   VIT_C
2   12     10
3   23     9
4   22     0

You just need to used a where clause

If your data looks like this Table [2]:

    COL1                  COL2
1   VIT_A                 VIT_C
2   12 23 22 0 32 0 11    10 9 0 21 0 26 76

Then you can use tranwrd function to replace ' ' with ',' then replace '0' with ''

Solution for table 1:

Data:

data have;
input VIT_A  VIT_C;
datalines;
12 10
23 9
22 0
0 21
3 20
0 26
11 76
;
run;

Create two new tables without any 0 values and add row number in as obs:

proc sql;
create table VIT_A as
select monotonic() as obs , VIT_A from have where VIT_A ne 0 ; 
create table VIT_C as
select monotonic() as obs ,VIT_C from have where VIT_C ne 0;
quit;

Merge the two tables:

data want;
merge VIT_A VIT_C;
by obs;
run;

Output:

 obs=1 VIT_A=12 VIT_C=10 
 obs=2 VIT_A=23 VIT_C=9 
 obs=3 VIT_A=22 VIT_C=21 
 obs=4 VIT_A=3 VIT_C=20 
 obs=5 VIT_A=11 VIT_C=26 
 obs=6 VIT_A=. VIT_C=76

Output

Upvotes: 1

Tom
Tom

Reputation: 51611

Just use a WHERE statement?

proc print data=have ;
  where vit_a ne 0 ;
run;

If you want to test multiple variables then make sure to combine them with the right logic.

where vit_a ne 0 AND vit_c ne 0;

Is different than

where vit_a ne 0 OR vit_c ne 0;

Upvotes: 0

momo1644
momo1644

Reputation: 1804

You can transpose your data, then select none zero values into a macro variable.

Assuming the values are in columns not in rows:

data have;
length string $100. ;
format string $100. ;
input string $ v1 v2 v3 v4 v5 v6 v7;
datalines;
VIT_A 12 23 22 0 32 0 11
VIT_C 10 9 0 21 0 26 76 
;
run;
proc transpose data=have out=trans name=value;
by string;
run;
proc sql noprint ;
select COL1 into :VIT_A separated by ' '  from trans where string="VIT_A" and COL1 <> 0;
select COL1 into :VIT_C separated by ' ' from trans where string="VIT_C" and COL1 <> 0;
quit;

Data Want;
string='VIT_A';
Values="&VIT_A.";
output;
string='VIT_C';
Values="&VIT_C.";
output;
run;

Output:

string=VIT_A Values=12 23 22 32 11
string=VIT_C Values=10 9 21 26 76

Upvotes: 0

Related Questions