Comparing multiple items

Question

I am trying to compare if a field has any value among a list of integers.( Ex: if MCC = 3001 || 3002 || 30004) in SPSS modeller. But the list of possible values is very long. Can anyone tell me how I can pass a list to it for comparing single values without using 'Or' / '||'.

Answer 1

This is what you are looking for:

If any(MCC, 3001, 3002, value3, value4, value5 ....) 

The condition is met if MCC contains any of the values listed after it.

Answer 2

The modeler CLEM expression is: member(@FIELD , [ int1 int2 int3 ])

with strings respectively: member(@FIELD , [ 'str1' 'str2' 'str3' ])

For more functions and related knowledge, here: https://www.ibm.com/support/knowledgecenter/en/SS3RA7_15.0.0/com.ibm.spss.modeler.help/clem_function_ref_comparison.htm

Answer 3

In SPSS Modeler you can use put all the values ( Ex: if MCC = 3001 || 3002 || 30004) in a Excel file or a text file.

And later, use a merge node to check if the values are in the data or not (Using the key variable MCC to compare the two datasets).

Answer 4

If the values are simply a continuous range then you can use RANGE:

COMPUTE MCC_List=RANGE(MCC, 3001, 3004)=1.

If there are a small number of discrete values that need to be exclude then you could use a combination of RANGE and ANY, like so:

COMPUTE MCC_List=RANGE(MCC, 3001, 3004)=1 and ANY(MCC,3003,3005)=0.

If your values are somewhat more random and discrete then an alternative way of achieving this, is to have the values stored as a look up table and then match on these values.

Below I manually enter these values in a dataset but typically you would save these in either a text/csv or Excel file and then you would read in that file.

This can be a good Data Management technique, particularly so if you have many variables where you have such definitions to derive. As you can store a separate file containing all these values without having to hard code them into the syntax and these files can simply be read in by the program and be updated if required.

(Although this technique does require a SORT CASES which may slow down the process if processing large datasets. Hence this type of technique is more commonly used in SQL. You could use SPSSs STAR JOIN which doesn't require a SORT but I'm not a big fan of that command myself).

Here is a quick demo:

DATA LIST LIST /ID MCC.
BEGIN DATA
1 4
2 5
3 1
4 6
5 7
END DATA.
DATASET NAME DS1.
SORT CASES BY MCC.

DATA LIST LIST /MCC_LIST MCC.
BEGIN DATA
1 1
1 2
1 3
1 4
1 5
END DATA.
DATASET NAME DS2.
SORT CASES BY MCC.

DATASET ACTIVATE DS1.
MATCH FILES FILE=* /TABLE=DS2 /BY MCC.
EXE.

Answer 5

The question was asked about Modeler, not Statistics.