how to create a dataset

[arpit]

I have a row data like

THE NAME OF COUNTRY IS INDIA
THE NAME OF COUNTRY IS AMERICA
THE NAME OF COUNTRY IS LONDON
THE NAME OF COUNTRY IS CANADA

I have to make a data set which is like

name
INDIA
AMERICA
LONDON
CANADA

Please help me Message was edited by: arpit

[sbb]

Use the SCAN function with -1 as the second argument to select the last word.

Scott Barry
SBBWorks, Inc.

Suggested Google advanced search argument, this topic / post:

scan function site:sas.com

[arpit]

The .csv file contain below data

THE NAME OF COUNTRY IS INDIA
THE NAME OF COUNTRY IS AMERICA
THE NAME OF COUNTRY IS LONDON
THE NAME OF COUNTRY IS CANADA



I have to create a dataset like

name
INDIA
AMERICA
LONDON
CANADA



can u explain me

[sbb]

From the data presented, it does not appear to be a "CSV" formatted file. What is unclear is whether the entire data-string is contained in one cell or multiple?

You can use SAS PROC IMPORT to read the external file -- this process will create a SAS database member/file for you. Then you can use a SAS DATA step with a SET to read your imported SAS file, and use a SAS character variable assignment statement. Code a LENGTH or ATTRIB statement to declare the new variable that will contain the sub-string value, and make use of the SCAN function, as shown below:

data _null_;
oldvar = "A B C D XXXXXXX';
length newvar $20;
newvar = scan(oldvar,-1,' ');
putlog _all_;
run;

Suggest using the Google searches below and review SAS-hosted documentation and supplemental technical / conference reference material.

Scott Barry
SBBWorks, Inc.

Suggested Google advanced search argument, this topic / post:

proc import csv introduction site:sas.com

data step programming introduction site:sas.com

data step scan function site:sas.com

[arpit]

Thanks for your kind information

But how it use to make a data step like

INDIA
AMERICA
LONDON
CANADA Message was edited by: arpit

[FloydNevseta]

There's a couple of ways you can do it. I'm using instream data in my code, but the data could be in a text file.

This is the method that Scott referred to in his post:

data method1;
input line $40.;
length country $8;
country = scan( line, -1 );
drop line;
datalines;
THE NAME OF COUNTRY IS INDIA
THE NAME OF COUNTRY IS AMERICA
THE NAME OF COUNTRY IS LONDON
THE NAME OF COUNTRY IS CANADA
;
run;


Another alternative:

data method2;
input @'THE NAME OF COUNTRY IS ' country $;
datalines;
THE NAME OF COUNTRY IS INDIA
THE NAME OF COUNTRY IS AMERICA
THE NAME OF COUNTRY IS LONDON
THE NAME OF COUNTRY IS CANADA
;
run;

[Ksharp]

Hi:
How about it:
[pre]
filename csv 'c:\test.csv';
data want;
infile csv length=len;
input whole $varying200. len;
country=scan(whole,-1);
drop whole;
run;
[/pre]



Ksharp