topic Re: Proc freq and alphanumeric variable in New SAS User

Proc freq and alphanumeric variable

femiajumobi1 — Mon, 19 Jun 2023 19:44:22 GMT

Hi. I need to generate frequencies of specific variables (TEM and DHA) which has alphanumeric values

TEM has: TEM, TEM-1, TEM-1(b), TEM-10, TEM-10(e)

DHA has: NEG, DHA-1, DHA-15,DHA-TYPE,DHA

Proc freq data=AMR;
Tables TEM DHA;
run;

The SAS codes showed only the following.

it appears it did not read the alphanumeric values.Kindly suggest what needs to be done. Thanks

Re: Proc freq and alphanumeric variable

Reeza — Mon, 19 Jun 2023 20:14:55 GMT

Back up a step and ensure your data is read in correctly. How did you import your data into SAS?

Re: Proc freq and alphanumeric variable

femiajumobi1 — Mon, 19 Jun 2023 20:22:10 GMT

I imported the data (in csv) with this code

proc import datafile = "....csv"
out = Add
dbms = csv
replace;
run;

Re: Proc freq and alphanumeric variable

Reeza — Mon, 19 Jun 2023 20:37:16 GMT

PROC IMPORT guesses at the lengths.

Try adding GUESSINGROWS=MAX to the code or write your own data step to read the data in correctly, with the correct lengths and then recheck it.

Re: Proc freq and alphanumeric variable

femiajumobi1 — Mon, 19 Jun 2023 21:45:55 GMT

I added "guessingrows=max" but it gave this response

There is unending running of string variables in the log

I changed to "guessingrows=15, but the frequencies omitted the alphanumeric variables still. Kindly suggest a way forward. Thank you.

Re: Proc freq and alphanumeric variable

Reeza — Mon, 19 Jun 2023 21:46:46 GMT

Other option - manually write a data step to import the data. There should be code in the log. Copy, paste that and then modify it as necessary.

Re: Proc freq and alphanumeric variable

Kurt_Bremser — Tue, 20 Jun 2023 06:12:21 GMT

NEVER use PROC IMPORT for csv files. You'll spend more time fixing stuff caused by the guessing than you need to write the DATA step yourself.

Re: Proc freq and alphanumeric variable

ballardw — Tue, 20 Jun 2023 15:16:50 GMT

The Log window full means that something about the file is generating lots of text. The most likely cause is something related to invalid data.

Question: did you ever open this CSV file in spreadsheet software? Did you SAVE it after opening with spreadsheet software? If the answer to both of these questions is YES then your file may have had the values changed by the spreadsheet and could be quite unreliable.

If the answer is know then open the CSV file in some text editing program like Notepad, copy the first 10 rows, then on the forum open a text box using the </> and paste the text. The text box is important to keep the forum from reformatting white space in text.

From what your picture shows I have feeling that your data file has many problems. Possibly related to line lengths and misaligned columns. I have a hard time believing that a column that would have value like "65-84 years" also has "<-0.06". That typically would indicate that a column, if the file is actually CSV, has 2 or more different types of values and that is not normal for a proper CSV file.

When you set Guessingrows to 15 that is even fewer rows than the default.

Re: Proc freq and alphanumeric variable

femiajumobi1 — Tue, 20 Jun 2023 16:14:12 GMT

Thanks. Writing the datastep as shown below does not help either.

%let path=C:\...;
options validvarname=v7;

proc import datafile="&path\....csv"
out=add
dbms=csv replace;
guessingrows=max;
run;

Re: Proc freq and alphanumeric variable

Reeza — Tue, 20 Jun 2023 16:20:29 GMT

The code you posted is not a data step, that's still proc import. A data step would look something like this:

data add;
infile "&path\...csv" dsd truncover;
informat var1 $8. var2 8. var3 8. var4 $12.;
input var1 var2 var3 var4;
run;

Here's a short tutorial on how to do so, and as mentioned previously PROC IMPORT puts the code it uses in the log. You can copy it and modify the type/format as needed for your data.

http://eta.health.usf.edu/publichealth/PHC6701/presentations/2013/readingExternalFilesSAS.pdf

@femiajumobi1 wrote:

Thanks. Writing the datastep as shown below does not help either.

%let path=C:\...;
options validvarname=v7;

proc import datafile="&path\....csv"
out=add
dbms=csv replace;
guessingrows=max;
run;

Re: Proc freq and alphanumeric variable

Kurt_Bremser — Tue, 20 Jun 2023 17:17:37 GMT

@femiajumobi1 wrote:

Thanks. Writing the datastep as shown below does not help either.

%let path=C:\...;
options validvarname=v7;

proc import datafile="&path\....csv"
out=add
dbms=csv replace;
guessingrows=max;
run;

This is NOT a DATA step.