BookmarkSubscribeRSS Feed
ajackson
Fluorite | Level 6

Hello,

 

I am requesting assistance on how to enter a query to find data that are flagged with errors within a dataset.

 

Thank you!

 

7 REPLIES 7
SASKiwi
PROC Star

You need to explain what you mean by "data that are flagged with errors within a dataset". What would be an example?

ajackson
Fluorite | Level 6

Hello,

 

I'm trying to figure out the total count of errors by error flag and quarter in my dataset. Would this be an if/then statement that I use? 

Tom
Super User Tom
Super User

@ajackson wrote:

Hello,

 

I'm trying to figure out the total count of errors by error flag and quarter in my dataset. Would this be an if/then statement that I use? 


Sound like you just want to run PROC FREQ.

So assuming you have a error flag variable name ERROR_FLAG and a variable named DATE that has date values you could probably just use the YYQ format to group by quarter.  So just do something like:

proc freq data=mydata ;
  tables date * error_flag;
  format date yyq6.;
run;

If there are a lot more distinct values of ERROR_FLAG than there are quarters you might want to reverse the two variables in the TABLES statement.  Or perhaps add the /LIST option to the TABLES statement.

Reeza
Super User

@ajackson wrote:

Hello,

 

I'm trying to figure out the total count of errors by error flag and quarter in my dataset. Would this be an if/then statement that I use? 


Depends a bit on your data structure. 

Assuming you have something like this:

 

Date Error_Flag
2022-10-23 Error Type 1
2022-10-24 Error Type 3

 

Then you can do something like this to get the results by quarter:

 

proc freq data=have; *have is the input data set;
table date*error_flag / list out=want;
format date yyq6.; *controls what level the errors are aggregated at;
*for monthly summary use yymmn6.;
run;

The results will be displayed and saved in a data set named want. 

ajackson
Fluorite | Level 6

Great. Thank you!

Reeza
Super User

What is an error?
If you're looking for data quality measures, that's a much bigger bag of worms.

 

https://www.lexjansen.com/mwsug/2010/stats/MWSUG-2010-182.pdf