Name the column in DataError for time zone values in every cursor

Part of #1212.

Summary

Time zone conversion failures raise DataError without the column name, and PandasCursor wraps them in OperationalError. Raise DataError naming the column in every cursor.

Facts

Measured on master ff4e468 (after #1224):

Cursor, S3FSCursor, and the fetch methods raise DataError for an unknown zone, an ambiguous local time, or a local time in a DST gap. The message names the value and the zone, but not the column.
PandasCursor raises OperationalError with the same message in whole reads.

Expected behavior

Each of these raises DataError with the column name, and for an unknown zone also the zone. Errors found by a chunked read raise when that chunk is read.

Years outside 1 to 9999 are out of scope (#1212): the errors that pandas, pyarrow, Polars, or Python raise for them pass through.

Follow the time zone errors of #1220 and simulate an older tzdata

The PandasCursor part is done in #1273: a timestamp with time zone value that PyAthena's conversion cannot convert, such as one with an ambiguous, nonexistent, or unknown zone, raises DataError naming the column, in whole and chunked reads, instead of an OperationalError that wraps it. When parse_dates, dtype, or converters is given to execute(), the default converter still runs per value as before, and an unknown zone raises OperationalError without the column name. The other cursors remain for this issue.