Fix _arrange_cols: stop mutating the caller's properties list

[thodson-usgs]

Summary

_arrange_cols did properties.append("geometry") and properties[properties.index("id")] = output_id in place. Calling any getter twice with the same properties=[...] list therefore caused that list to grow unboundedly and have id rewritten across calls — a real action-at-a-distance defect for any user code that re-uses a properties list across requests.

This PR takes a local copy of the list and mutates that. The caller's list is untouched.

Minimal reproducible example

A common pattern: define a properties list once and reuse it for two different getters.

from dataretrieval.waterdata import get_daily, get_continuous

cols = ["id", "monitoring_location_id", "value", "time"]

# Call 1 - get_daily mutates `cols` in place
get_daily(
    monitoring_location_id="USGS-05427718",
    parameter_code="00060",
    time="2025-01-01/2025-01-02",
    properties=cols,
)
# cols is now ['daily_id', 'monitoring_location_id', 'value', 'time', 'geometry']

# Call 2 - same list, different getter, hits the live API with the
# already-mutated names
get_continuous(
    monitoring_location_id="USGS-06904500",
    parameter_code="00065",
    time="2025-01-01T00:00:00Z/2025-01-01T01:00:00Z",
    properties=cols,
)
# RuntimeError: 400: InvalidParameterValue. unknown properties specified.

The 'id' -> 'daily_id' rewrite from call 1 leaks into the wire request for get_continuous, which doesn't recognize daily_id as a valid property and rejects the entire request. With this PR, the same code succeeds — cols is byte-for-byte unchanged after each call. The same defect surfaces in three more everyday patterns (loop over sites with shared list, get_continuous -> get_daily, get_daily -> get_monitoring_locations); all four fail with 400 against the live API on main and all four succeed on this branch.

Test plan

• Three new regression tests in tests/waterdata_utils_test.py: caller list is unchanged after two consecutive calls; 'id' still resolves to the output_id column in the result; geometry is still tacked on when present in the response.
• Existing tests in the file still pass.
• Full local test suite passes (with API_USGS_PAT set).
• Verified live against api.waterdata.usgs.gov that the four cross-getter / loop scenarios above fail with 400 on main and succeed on this branch.

Related PRs

Other open PRs in this bug-review series that touch dataretrieval/waterdata/utils.py (different functions, no functional conflicts):

• Fix _format_api_dates: parse ISO 8601 datetimes #247 — _format_api_dates accept ISO 8601.
• waterdata: get_stats_data preserve geometry across continuation pages #255 — get_stats_data preserve geometry across continuation pages.
• Make _handle_stats_nesting tolerant of missing drop columns #257 — _handle_stats_nesting tolerate missing drop columns.

[Copilot]

Pull request overview

This PR fixes an action-at-a-distance bug in dataretrieval.waterdata.utils._arrange_cols where the function mutated the caller-provided properties list in-place, causing surprising behavior when the same list instance is reused across multiple requests.

Changes:

• Stop mutating the caller’s properties list by copying it to a local list before applying id/geometry adjustments.
• Add regression/behavior tests covering non-mutation, id→output_id behavior, and retention of geometry when present.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
dataretrieval/waterdata/utils.py	Copies properties before mutation to prevent side effects across calls.
tests/waterdata_utils_test.py	Adds regression + behavior tests validating the fixed _arrange_cols semantics.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated no new comments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.