fix combine_merged_cells when using from a range that doesn't start at A1

[alifeee]

closes #1330

add start_row_index and start_col_index to utils.combined_merge_values so that ranges that don't start at col 0, row 0 (i.e., A1) can have their merged cells merged.

n.b., not starting from the top left introduces more issues with combine_merged_cells -> #1334

This code is added to worksheet.get_values. This is to get the start_row_index and start_col_index from the requested range.

gspread/gspread/worksheet.py

Lines 495 to 503 in 0e698de

	grid_range = a1_range_to_grid_range(
	get_a1_from_absolute_range(range_name),
	)
	return combined_merge_values(
	worksheet_metadata=worksheet_meta,
	values=vals,
	start_row_index=grid_range.get("startRowIndex", 0),
	start_col_index=grid_range.get("startColumnIndex", 0),
	)

• This may fail if a named range is used (see A1 notation docs) - add tests? or do not allow combine_merged_cells if a named range is used? since we don't know the start index?

[alifeee]

@lavigne958 once this is approved and merged, we can release 5.12.1.

This release should be the last before 6.0.0, so without another (5.12.2 e.g.), this will also the last chance to get an updated README onto the PyPi (the readme on the GitHub will always stay up to date)

[lavigne958]

Sure I need to take a look properly when I have time. This week should do.

[lavigne958]

• This may fail if a named range is used (see A1 notation docs) - add tests? or do not allow combine_merged_cells if a named range is used? since we don't know the start index?

When we request: same value for merged cells, we make a second call to the spreadsheet right ? to pull all the merges.
Can we extract the coordinates of the named range at the same time ? 🤔

[alifeee]

When we request: same value for merged cells, we make a second call to the spreadsheet right ? to pull all the merges.
Can we extract the coordinates of the named range at the same time ? 🤔

Yes. In fact, we already request the spreadsheet metadata to get the merges:

gspread/gspread/worksheet.py

Line 490 in bcac8cb

spreadsheet_meta = self.spreadsheet.fetch_sheet_metadata()

So, I have reused this metadata to also find the grid_range of a named range:

gspread/gspread/worksheet.py

Lines 496 to 507 in bcac8cb

	# deal with named ranges
	named_ranges = spreadsheet_meta.get("namedRanges", [])
	# if there is a named range with the name range_name
	if any(
	range_name == ss_namedRange["name"]
	for ss_namedRange in named_ranges
	if ss_namedRange.get("name")
	):
	ss_named_range = finditem(
	lambda x: x["name"] == range_name, named_ranges
	)
	grid_range = ss_named_range.get("range", {})

and other range types are in a different branch

gspread/gspread/worksheet.py

Lines 508 to 514 in bcac8cb

	# norrmal range_name, i.e., A1:B2
	elif range_name is not None:
	a1 = get_a1_from_absolute_range(range_name)
	grid_range = a1_range_to_grid_range(a1)
	# no range_name, i.e., all values
	else:
	grid_range = worksheet_meta.get("basicFilter", {}).get("range", {})

This may be too much logic for get_values and you might prefer it be offloaded to a utils function. It also is only implemented for get_values, as get and get_values have not yet been combined (happens in v6.0.0)

[lavigne958]

let's get release 5.12.1 out

[alifeee]

agreed. Does that include a fix for #1354?

[lavigne958]

I dont't think we should include it, it still needs a bit more thought about the resolution of the problem.
This is release has already quiet a bit of changes to provide.
In my opinion: we should merge this one and release.

#1354 can be addressed right after.