get_table_data(): "False" string incorrectly converts to True (bool conversion pitfall)

[DavidNeedsHiking]

Summary

When reading DataFrame data via get_table_data() or query_table_data(), boolean values are returned as strings ("True"/"False"). Converting these back to Python booleans using bool() or pd.Series.astype(bool) incorrectly treats "False" as True because any non-empty string is truthy in Python.

Steps to Reproduce

from nisystemlink.clients.dataframe import DataFrameClient

client = DataFrameClient()
result = client.get_table_data(table_id)

# result.frame.data contains: [["1", "False"], ["2", "True"]]
# Attempting to convert:
bool("False")  # Returns True! (incorrect)

Expected Behavior

Boolean string values should convert correctly: "False" → False, "True" → True

Suggested Fix

Add a helper method or conversion utility that uses explicit mapping instead of bool():

_bool_map = {'true': True, '1': True, 'false': False, '0': False}
series.str.strip().str.lower().map(_bool_map)

1. Short-term fix: Add a utility function like frame_to_pandas(result.frame, dtypes={...}) that handles boolean (and other type) conversions safely

2. Long-term fix: Return Arrow/binary data for read operations too, matching the write format — eliminating the JSON string conversion entirely

Environment

Python 3.x
nisystemlink-clients (latest)

Related

This asymmetry exists because writes use Arrow binary format (types preserved) while reads use JSON (all strings). A frame_to_pandas() helper in the SDK would save users from this pitfall.