For pandas 1.3.2, tests currently fail because objects are converted to floats automatically in pandas 1.1.5 (which we use for testing) when possible but not in 1.3.2. This causes sr.str to raise an error in 1.3.2, this is fixed by changing the sr.str call to an explicit sr == "" in the infer_dtypes function
For pandas 1.3.2, tests currently fail because objects are converted to floats automatically in pandas 1.1.5 (which we use for testing) when possible but not in 1.3.2. This causes sr.str to raise an error in 1.3.2, this is fixed by changing the sr.str call to an explicit sr == "" in the
infer_dtypes
function