pandas column-type overwrite

[isichei]

Allow option to overwrite standard conversion e.g.

{
  "int": "character",
  "long": "character"
}

Should also enable specific column type overwrite:

{
  "col2": "int"
}

☝️ First thinking is to always use agnostic meta data types to make it easier for the user to understand. However there is probably more flexibility in pandas datatypes so maybe there should be the option for both.

[isichei]

This issue will be solved / depreciated by issue #15