Meaningful zeroes on your substantive variables (such as gender) are not a problem, and are not what this error message refers to.

If you have declared missing values (such as 0, 9, or -99) then those should not be a problem either (unless they are in the computation of the weighting variable -- see below on that issue). SPSS will exclude cases with appropriately declared missing data. There are various ways you can specify how such cases are excluded (listwise deletion, pairwise deletion), but if you don't specify then SPSS will use its default method of exclusion. Again, that is not what this error message is about.

Rather, this error message refers only to whatever variable you have specified in the WEIGHT command for SPSS to use in weighting the cases. Ordinarily, you would not want the weighting variable to have any cases assigned a value of zero. As the error message indicates, all such cases would be excluded from your analyses. If you want to drop some cases from a particular analysis, it's usually better to use SELECT IF for that purpose, not zero weights.

I suggest you run FREQUENCIES on your weighting variable to determine how many zero (or blank, or negative) values the error message is flagging for you. Then look into the data file and determine what characteristics these cases share. It sounds like something has gone wrong in the coding or calculation of the weighting variable that has given it zero, blank, or negative values that were not intended.

If you did a COMPUTE to create that weighting variable from other variables in the file, and some of those other variables had any declared missing values, then the computed weighting variable would also be missing (blank) for every case that had any missing data. That's not good; every case needs a positive weight. The WEIGHT command would read those blanks as zero, and give you this error message.

Or perhaps your data file has one or more blank rows. The WEIGHT command would read each of those rows as a case with zero weight, so again you would get the error message. This last possibility is the easiest one to fix -- if that's the whole problem, you can either ignore the error message, or delete the blank rows from your data file and you will stop getting it.

Hope this helps.

The question was asked on IBM and here is their answer.

https://www.ibm.com/support/pages/receiving-warning-3211-when-performing-analysis

If it doesn't help you, try contacted IBM directly.

I think this link can help you

https://www.ibm.com/support/pages/receiving-warning-3211-when-performing-analysis