input.Makefile: Paths: $(datasrc): Remove any "." prefix from the subdir name. The "." prefix allows a subdir to be hidden from the normal import process.

db_xml.py: put_table(): Allow caller to specify custom partition_size

tnrs.py: tnrs_request(): Return the CSV stream directly instead of reading it into a string

tnrs.py: tnrs_request(): Moved CSV-download-specific functionality from do_request() to the Download section

inputs/import.stats.xls: Updated import times

tnrs.py: tnrs_request(): Return the response instead of printing it to stdout

schemas/py_functions.sql: _namePart(): Fixed bug where it was returning the empty string instead of NULL

sql_io.py: import_csv(): Documented that sql.truncate() MUST be run so that the rows will be stored in inserted order, and the row_num added after import will match up with the CSV's row order

sql.py: add_row_num(): Add distinguishing comment to ADD COLUMN statement so that it will be cached. The distinguishing comment is required because sometimes column names are truncated, leading to unwanted collisions with previously-cached ADD COLUMN statements. It provides a way of distinguishing the full column name behind a particular ADD COLUMN statement.

sql_io.py: import_csv(): Free memory used by deleted rows from any failed import. Documented that sql.create_table() is not rolled back if the import fails, but instead is cached, and will not be re-run if the import is retried.