error in merging : Multiple lengths were specified for the BY variable

[archibald]

Hello,

I have two datasets that I need to merge using the ID variable (Inmatesid)

DATA X

Inmatesid	Race
03-000011ID	2
30-000211ID	3
193-004000ID	1
1903-022222ID	1
1903-001244ID	1
1903-001246ID	1
1903-1265790ID	5
1903-123457ID	2
1903-123456ID	2
1903-000055ID	4
1903-122000	2

DATA Y

Inmatesid	Tatoo	Security level
506-1903-000011	Yes	3
506-1903-000211	No	.
506-1903-004000	Yes	1
506-1903-022222	Yes	2
506-1903-001244	Yes	4
506-1903-001246	No	.
506-1903-126579	Yes	2
506-1903-123457	No	2
506-1903-123456	Yes	2
506-1903-000055	Yes	3
506-1903-002222	Yes	1
506-1903-011123	No	5
506-1903-122000	Yes	5

My first task  was  to standardize  inmate ID  in both datasets so that

1903-000055ID =>1903-000055    (data x)

506-1903-000055 =>1903-000055  (data y)

1) for data  x,  I used inmatesid=compress(inmatesid, ' ', 'I');

2) data y,  i used   Inmatesid=trim(left(substr(Inmatesid, index(Inmates id, '-')+1)));

 the first issue I have is that  there are some typos for inmatesid (for example  03-000011ID) in data x,  which  does not work well with that above code. so I manually tried to fix  the ID's that were not transformed properly; I will  be curious to know  how to  do this using a  code .

Then I tried to merge the two data sets but I am getting this error. Does anybody know how to debbug this error. Thanks

WARNING: Multiple lengths were specified for the BY variable Inmatesid by input data sets and LENGTH,

         FORMAT, INFORMAT, or ATTRIB statements. This may cause unexpected results.

WARNING: Multiple lengths were specified for the variable Inmatesid by input

         data set(s). This may cause truncation of data

[ballardw · Accepted solution]

First, that is not an error. The cause is if you look at the Length property of the variable you will find that the two are different.

The solution to removing the warning is to ensure like variables have the same length before combining data sets.

There is a possibility that things won't work right because the "shorter" values may not match a longer one.

Example:

Id = 123456 in one set and in the other set

id = 1234567 but are supposed to be the "same". Since they are not identical then the BY treats them as different records.

Also the first dataset listed will set the length of the output variable. So in the case above the ID of 1234567 may get truncated to 123456 and then you spend time trying to figure out why the sets apparently didn't merge properly.

[archibald]

Makes sense! thank you!