fetchdf() and fetchnumpy() do not work for all data types

[evangelos-dimitriadis]

Describe the bug
fetchall() will return all the data by rows and it works fine.
However fetchnumpy() and fetchdf() work only with INTEGERS, FLOATS and BOOLEAN.
I tried UUID, STRING, VARCHAR, CHAR and all return the same error.

To Reproduce

import monetdbe

if __name__ == "__main__":
    conn = monetdbe.connect(':memory:')
    if not conn:
        print('Could not access the memory')

    c = conn.cursor()
    c.execute("CREATE TABLE charachters(j STRING);")
    c.execute("INSERT INTO charachters VALUES ('one'), ('two'), ('three');")
    c.execute("SELECT * FROM charachters;")

    for r in c.fetchall():
        print(r)

    c.execute("SELECT * FROM charachters")
    print(c.fetchdf())
    c.execute("SELECT * FROM charachters")
    print(c.fetchnumpy())

    c.close()
    conn.close()

Expected behavior
Print the selected fields

Software versions

• OS and version: Ubuntu 19.04
• Python version: 3.8
• MonetDBe-Python version: 0.7

Additional context

The error message is:

ValueError: itemsize cannot be zero in type

[gijzelaerr]

working on this in https://github.com/MonetDBSolutions/MonetDBe-Python/tree/issue62_complex_numpy_columns

[gijzelaerr]

This should be resolved in the PR and will be included in the 0.8 release