Select data with only 4 decimals

[podarum]

Hi,  I have a dataset that is interest rate, most of the data is in a 2 decimal format (eg. 2.45), but some rows have a 4 decimals (eg. 2.4678) format.  How do I chose the data that has only the 4 decimals?  Thanks

[novinosrin · Accepted solution]

data have;
 k=2.4678;
 output;
 k=2.34;
 output;
 k=2;
 output;
 k=2.4450;
 output;
 k=2.4400;
run;

data want;
 set have;
 if mod(k*1e4,10);
run;

 /*or where*/

 where mod(k*1e4,10);

[PaigeMiller]

The solution from @novinosrin fails if k=2.46789.

How about this:

data want;
	set have;
	ychar=put(k,best12.);
	if length(strip(scan(ychar,2,'.')))=4;
run;

 And @Tom makes a good point too... 

[novinosrin]

Yes, that's correct. The solution assumes based upon OP wrote "most of the data is in a 2 decimal format (eg. 2.45), but some rows have a 4 decimals (eg. 2.4678) format."

[podarum]

Thanks everyone, the answers are great

[Tom]

---

@podarum wrote:

Hi,  I have a dataset that is interest rate, most of the data is in a 2 decimal format (eg. 2.45), but some rows have a 4 decimals (eg. 2.4678) format.  How do I chose the data that has only the 4 decimals?  Thanks

---

If the field is not a character string then you don't.

How can you tell the difference between a value that was original written as '2.50' and a value written as '2.500'?  They are the same number. 

[Reeza]

What's your original data source? Can you read it in as character first?
This would solve your problem, technically....you may still have logic issues.

[ChrisNZ]

Another way:

 if length( scan( put(VAR,best32. -l), 2, '. ' ) ) = 4;