I am trying to do DESeq2 on the dataset GSE202203. I downloaded the data from the server using GSE202203_RawCounts_gene_3207 link. But when I opened the data, it was in fraction number as follows:

How do I effectively convert the values compatible for DESeq2 analysis?

Is this way correct? df[,-1] <-round(df[,-1],0)

Presumably those are the counts generated by tximport after importing the data. While it's not ideal (we don't know about gene length per sample), you can round these and load into DESeq2 with DESeqDataSetFromMatrix. They should have instead posted the lengths as well as the counts (which tximport keeps track of, but here was discarded).