topic Re: Reading Mixed Raw Data in SAS Procedures

Reading Mixed Raw Data

David_S — Fri, 05 Dec 2014 06:34:10 GMT

**Fictional Content used**

Hello,

I'm trying to read in a staggered raw data file with various column widths but you see that I cannot successfully use a pointer

(i.e. @19 or +2) since the Employee_Name variable isn't a fixed column. How do I best achieve results such that they appear as expected?

RAW DATA

(Level1) Chief Executive Officer (Anthony Miller ) $433,800

(Level2) Chief Sales Officer (Harry Highpoint ) $243,190

(Level2) Chief Financial Officer (Max Crown ) $268,455

EXPECTED

Level Job_Title Employee_Name Salary

1 Chief Executive Officer Anthony Miller $433,800

2 Chief Sales Officer Harry Highpoint $243,190

2 Chief Financial Officer Max Crown $268,455

My code is as follows:

Work.Unicorn;

infile "rawin.dat" missover truncover;

length Level $ 8.;

input Level @;

if Level="(Level1)" then

input @10 Job_Title $25. @35 Employee_Name $18. @108 Salary dollar8.;

else if Level="(Level2)" then

input Job_Title $ +2;

run;

Thank you for any assistance,

David

Re: Reading Mixed Raw Data

Patrick — Fri, 05 Dec 2014 07:00:05 GMT

You will need to get the informats right - but else below code should give you the idea.

data want;

infile datalines truncover dlm='()' dsd;

input dummy :$1 (Level Job_Title Employee_Name Salary) (:$20.);

datalines;

(Level1) Chief Executive Officer (Anthony Miller ) $433,800

(Level2) Chief Sales Officer (Harry Highpoint ) $243,190

(Level2) Chief Financial Officer (Max Crown ) $268,455

run;

Re: Reading Mixed Raw Data

Reeza — Fri, 05 Dec 2014 07:01:30 GMT

I would read the first portion as a single text field and then parse it using the scan function with the () as delimiters.

Re: Reading Mixed Raw Data

David_S — Fri, 05 Dec 2014 20:28:38 GMT

Thank you all.

I overlooked using () as delimeters. Thanks for the info.

David

Re: Reading Mixed Raw Data

Patrick — Fri, 05 Dec 2014 23:51:37 GMT

Then please mark the answer which is closest to what you wanted as "correct" so that this question gets "ticked off" as answered.