Interpret all columns as TEXT data type

[tsundara]

Is it possible to interpret all columns as TEXT datatype (through a flag maybe?)
I think the columns values are sampled and then column datatype is guessed. If there is an incompatible value in some row, then that row seems to be skipped. So instead, is it possible I load everything as TEXT data type?
I just need the data in some format(for data comparison purposes), But all csv data must go into the table without skipping rows.

Thanks
Thyag

[ericlentz]

Similarly, I have item numbers in the format, 00645155909931 and it reads it as an integer datatype which converts it to 643169607033.

[tsibley]

I also have this issue. I thought specifying --shape "csv_col:sql_col(text),…" for each column would work, but the text type isn't passed into pd.read_csv's dtype parameter.

[dannguyen]

It'd be great if there were an option akin to csvkit's csvsql --no-inference to force every column to be just TEXT, for imports in which there are just too many columns to easily specify the --shape flag:

https://csvkit.readthedocs.io/en/1.0.2/scripts/csvsql.html

[simonw]

This feature has now been released in csvs-to-sqlite 1.1