Python 将字符串转换为 object

Question

I am using pandas/sql alchemy to write a table to an sql database.

Here is some sample code below:

df.to_sql('table', connection, schema=None, dtype={
    'id': sa.types.VARCHAR(length=100),
    'value1': sa.types.Float(),
    'value2': sa.types.Float()})

I have a csv file where the data type for each column is defined, I would like to be able to read the csv file using pandas and then convert the table to a dictionary which would be passed to the to_sql() method.

If I were to do that directly, the data type, for example sa.types.VARCHAR(length=100) would appear as a string. How can I convert this string so that it can be passed as an appropriate data type.

To clarify if i read the data from a csv table, the dtype dictionary will look like this:

{'id': 'sa.types.VARCHAR(length=100)'}

instead of the current:

{'id': sa.types.VARCHAR(length=100)}

Answer 1

This solution is hacky and I have to believe there's a better way if we had all of the information, but short of that, here goes.

I would create a dictionary similar to:

type_dict = {
   'VARCHAR' : sa.types.VARCHAR,
   'Float'   : sa.types.Float, 
   ... }

Then, break up that string to get the main info you're looking for (I'm sure there's a better regex way to do this, but it's past my bedtime):

dtype = string.split('(')[0].split('.')[-1]
var   = string.split('(')[1].split('=').split(')')[0]

Then:

sql_dict = { 'id' : type_dict[dtype](var) }