Joining multiple tables with different keys using hash object

[subhani4]

i have data as below 

data class;
input Name$ Sex$ Age Height Weight;
infile datalines dlm=' ' dsd;
datalines;
Alfred M 14 69.0 112.5
Alice F 13 56.5 84.0
Barbara F 13 65.3 98.0
Carol F 14 62.8 102.5
Henry M 14 63.5 102.5
;

data zipcodes;
input name $ zipcode;
infile datalines dlm=' ' dsd;
datalines;
Alfred 91940
Alice 75008
Barbara 92330
Carol 94150
Henry 93140
;

data cities;
input zipcode City :$30. Department :$30.;
infile datalines dlm=' ' dsd;
datalines;
75008 Paris 8è Paris
91940 Les Ulis Essonne
92330 Sceaux Hauts-de-Seine
93140 Bondy Seine-Saint-Denis
94150 Rungis Val-de-Marne
;

i  wand the output as  

Name	ZipCode	City	Department
Alfred	91940	Les Ulis	Essonne
Alice	75008	Paris 8è	Paris
Barbara	92330	Sceaux	Hauts-de-Seine
Carol	94150	Rungis	Val-de-Marne
Henry	93140	Bondy	Seine-Saint-Denis

i have coded as below but not able to make it happen:

data need(keep=name zipcode city department);

if 0 then set zipcodes cities;
if _n_=1 then do;
dcl hash z(dataset:'zipcodes');
z.definekey('name','zipcode');
z.definedata(all:'Y');
z.definedone();
dcl hash c(dataset:'cities');
c.definekey('zipcode');
c.definedata(all:'Y');
c.definedone();
end;
do until(eof);
set class end=eof;
if z.find(key:name,key:zipcode)=0 and c.find(key:zipcode)=0;
end;
stop;

[Reeza]

What does not able to make it happen mean? If you have errors please include them in the post.

[Sajid01]

Hello @subhani4 

First make the cities table comma delimited. I believe your objective is to join the datasets to obtain an output as you desired.
Second it is easy to use proc sql to make the join. Please see the code below .

Information from zipcode and cities table is sufficient to produce the desired output.

data class;
input Name$ Sex$ Age Height Weight;
infile datalines dlm=' ' dsd;
datalines;
Alfred M 14 69.0 112.5
Alice F 13 56.5 84.0
Barbara F 13 65.3 98.0
Carol F 14 62.8 102.5
Henry M 14 63.5 102.5
;

data zipcodes;
input name $ zipcode;
infile datalines dlm=' ' dsd;
datalines;
Alfred 91940
Alice 75008
Barbara 92330
Carol 94150
Henry 93140
;

data cities;
input zipcode City :$30. Department :$30.;
infile datalines dlm=',' dsd;
datalines;
75008, Paris 8è ,Paris
91940, Les Ulis, Essonne
92330,Sceaux, Hauts-de-Seine
93140, Bondy , Seine-Saint-Denis
94150, Rungis, Val-de-Marne
;
proc sql;
create table want as
select z.name, z.zipcode, ci.City Format $30. length=30 , Ci.Department Format $30. length=30
from zipcodes z
Left join Cities ci ON z.zipcode=ci.zipcode
order by Name;
quit;
proc print;
run;

The output will be as follows

[Kurt_Bremser]

do until(eof);
set class end=eof;
if z.find(key:name,key:zipcode)=0 and c.find(key:zipcode)=0;
end;
stop;

Since there is no OUTPUT statement in your loop, and the STOP statement terminates the data step before the implicit output is done, nothing is written to the dataset.

Use the "natural loop" of the data step instead:

data need(keep=name zipcode city department);
set class;
if 0 then set zipcodes cities;
if _n_=1 then do;
  length zipcode 8 city department $30; 
  dcl hash z(dataset:'zipcodes');
  z.definekey('name');
  z.definedata('zipcode');
  z.definedone();
  dcl hash c(dataset:'cities');
  c.definekey('zipcode');
  c.definedata('city','department');
  c.definedone();
end;
if z.find()=0;
if c.find()=0;
run;

Note that I adapted the hash definitions, as dataset class does not have a zipcode variable. Therefore it is also necessary to split the subsetting IFs, to make sure that the second FIND() is executed after the first one that retrieves the zipcode.

The code is not tested in any way, as I am posting from my tablet.