fix(tests): use correct compact table row format in std_dev and variance tests

[benbellick]

The compact aggregate test format uses each inner parenthesis group as a row. The std_dev and variance test files were using a flat list of values in a single group, e.g. ((1.0, 2.0, 3.0)), which parses as 1 row with 3 columns rather than 3 rows with 1 column.

This changes them to the correct per-row format: ((1.0), (2.0), (3.0)).

This format is documented here (this is format 2):

substrait/tests/README.md

Lines 71 to 84 in 87d73b6

	Aggregate test cases support 3 formats:
	1. **Single Argument**: The test case for an aggregate function with single argument as a column in a table. The table is defined in the test case.
	```code
	sum((1, 2, 3, 4, 5)::i8) = 15::i8 # addition of two numbers
	```
	2. **Multiple Columns Compact**: The test case for an aggregate function with on one or more columns of a table as argument. The table is defined before the function name, in the same line as the testcase.
	```code
	((20, 20), (-3, -3), (1, 1), (10,10), (5,5)) corr(col0::fp32, col1::fp32) = 1::fp64
	```
	3. **Multiple Columns Verbose**: The test case for an aggregate function with one or more columns of a table as arguments. The table is defined before the function name. The function arguments refer to the columns in a table.
	```code
	DEFINE t1(fp32, fp32) = ((20, 20), (-3, -3), (1, 1), (10,10), (5,5))
	corr(t1.col0, t1.col1) = 1::fp64
	```

This came up in some work on substrait-go.

(I opened #1042 to discuss how the inline-format 1 above could be used in the future. We would have to figure out how to deal with intermixing enum args.)

🤖 Generated with Claude Code

---

This change is 

[nielspardon]

LGTM, thanks for catching this