While doing the gene mapping from Affymatrix ids, I found out that there are multiple affymetrix IDs corresponding to a single gene name for a sample. My work revolves around gene names only. Please suggest the appropriate step to deal with this situation. The code is as follows:

require("biomaRt")
mart <- useMart("ENSEMBL_MART_ENSEMBL")
mart <- useDataset("hsapiens_gene_ensembl", mart)
annotLookup <- getBM(
  mart=mart,
  attributes=c(
    "affy_hg_u133_plus_2",
    "ensembl_gene_id",
    "gene_biotype",
    "external_gene_name"),
  filter = "affy_hg_u133_plus_2",
  values = rownames(exprs(gset))[1:dim(exprs(gset))[1]], uniqueRows=TRUE)

A sample output snap is attached for reference.

Thank you.

How you deal with duplicate measurements is up to you as the analyst. I would caution against taking advice from random strangers on the internet who know nothing about your study. That said, there are several things that people have done over the years, from taking the average (using e.g. avereps from limma), or choosing the 'best one' or simply accepting that there are multiple probesets that may interrogate a given gene.

I will point out that the Affymetrix ID is meant to provide some evidence as to the quality of the probeset. Anything that is XXX_at is thought to bind only to the gene of interest, and XXX_s_at may bind to multiple genes in a family, and XXX_x_at is even more speculative, possibly binding to unrelated genes. I forget the exact details, having not analyzed Affy data for years now. You might be able to track that information down somewhere, but after Fisher Scientific bought Affymetrix, the level of support seems to have been reduced significantly.