Is this single-cell RNAseq or non-single-cell RNAseq?

For single-cell, you typically don't do standard "differential gene expression analysis" and you should follow the typical tools used for single-cell analysis.

For non-single-cell, you should figure out why you're getting decimals. If this is because you're using a kallisto/RSEM-type of quantification, then you can use tximport before doing a standard DESeq2-type analysis.

Finally, you should find out how your "matrix" was generated. If you can't figure that out, then how am I supposed to figure it out and reproduce your analysis after you publish your paper in a journal and upload your sequencing reads?