schema_sample_rows=0 results in a table filled with null values

[niekrongen]

When reading a excel file and setting the schema_sample_rows to 0 results in a table of the correct height, but all values are set to null. If the schema_sample_rows is set to n and the first n values in a column are empty, then the remainder of the values in this column are also filled with null values. This means that the values that are present in this column, are replaced with null values. This is causing a loss of data, while this can be fixed by defaulting to a column dtype of type string.

For example, xlsx2csv has this also as default.