Serialize BigDecimal manually to match Arrow's Decimal128 format

[fabubaker]

What changes were proposed in this pull request?

Serializes Prop::Decimal, which internally holds a BigDecimal, to match the Decimal128 format used in Arrow.

Additionally, modifies column_as_array to return a Result.

Why are the changes needed?

parquet_tests::node_temp_props_decimal currently fails due to Prop::Decimal being serialized incorrectly using scientific notation. serde_arrow expects floats to be in standard notation.

Furthermore, having column_as_array (and its callers) return a Result allows prop serialization errors to bubble up properly. Currently for props that can't be serialized, errors are swallowed using .ok and return empty columns.

Does this PR introduce any user-facing change? If yes is this documented?

No.

How was this patch tested?

This patch fixes a failing test.

Are there any further changes required?

No.

[github-actions]

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'Rust Benchmark'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 2.

Benchmark suite	Current: 478d4c6	Previous: 9823ef7	Ratio
lotr_graph/num_edges	5 ns/iter (± 0)	0 ns/iter (± 0)	+∞
lotr_graph/num_nodes	5 ns/iter (± 0)	1 ns/iter (± 0)	5
lotr_graph/iterate_exploded_edges	603149 ns/iter (± 3628)	285473 ns/iter (± 2920)	2.11
lotr_graph/graph_latest	3 ns/iter (± 0)	0 ns/iter (± 0)	+∞
lotr_graph_materialise/materialize	8156928 ns/iter (± 64618)	1564816 ns/iter (± 35303)	5.21
lotr_graph_window_100/num_nodes	14 ns/iter (± 0)	5 ns/iter (± 0)	2.80
lotr_graph_window_100/iterate_exploded_edges	775785 ns/iter (± 1577)	325242 ns/iter (± 847)	2.39
lotr_graph_window_100_materialise/materialize	8549448 ns/iter (± 35869)	1669150 ns/iter (± 10700)	5.12
lotr_graph_window_10/has_edge_existing	211 ns/iter (± 5)	97 ns/iter (± 11)	2.18
lotr_graph_window_10/has_node_existing	153 ns/iter (± 8)	62 ns/iter (± 11)	2.47
lotr_graph_window_10/iterate nodes	33216 ns/iter (± 90)	11339 ns/iter (± 40)	2.93
lotr_graph_window_10/iterate edges	99624 ns/iter (± 400)	48684 ns/iter (± 211)	2.05
lotr_graph_window_10/iterate_exploded_edges	402719 ns/iter (± 1745)	155788 ns/iter (± 1001)	2.59
lotr_graph_window_10_materialise/materialize	3679340 ns/iter (± 8390)	971980 ns/iter (± 4278)	3.79
lotr_graph_subgraph_10pc_materialise/materialize	2068365 ns/iter (± 18520)	334634 ns/iter (± 1287)	6.18
lotr_graph_subgraph_10pc_windowed/has_node_existing	158 ns/iter (± 8)	62 ns/iter (± 14)	2.55
lotr_graph_subgraph_10pc_windowed/iterate nodes	5439 ns/iter (± 97)	1365 ns/iter (± 3)	3.98
lotr_graph_subgraph_10pc_windowed_materialise/materialize	1257737 ns/iter (± 14141)	230399 ns/iter (± 2617)	5.46
lotr_graph_window_50_layered/num_edges_temporal	149993 ns/iter (± 2450)	70121 ns/iter (± 7586)	2.14
lotr_graph_window_50_layered/has_edge_existing	210 ns/iter (± 6)	104 ns/iter (± 7)	2.02
lotr_graph_window_50_layered/has_node_existing	507 ns/iter (± 16)	129 ns/iter (± 12)	3.93
lotr_graph_window_50_layered/iterate nodes	77449 ns/iter (± 239)	19308 ns/iter (± 47)	4.01
lotr_graph_window_50_layered/iterate edges	200924 ns/iter (± 537)	83616 ns/iter (± 1318)	2.40
lotr_graph_window_50_layered/graph_latest	77935 ns/iter (± 1497)	36649 ns/iter (± 916)	2.13
lotr_graph_window_50_layered_materialise/materialize	32223075 ns/iter (± 154269)	3488825 ns/iter (± 24948)	9.24
lotr_graph_persistent_window_50_layered/num_edges_temporal	591852 ns/iter (± 5325)	192686 ns/iter (± 1569)	3.07
lotr_graph_persistent_window_50_layered/has_edge_existing	227 ns/iter (± 106)	109 ns/iter (± 23)	2.08
lotr_graph_persistent_window_50_layered/has_node_existing	525 ns/iter (± 260)	174 ns/iter (± 83)	3.02
lotr_graph_persistent_window_50_layered/iterate nodes	98797 ns/iter (± 346)	35886 ns/iter (± 191)	2.75
lotr_graph_persistent_window_50_layered/iterate edges	184923 ns/iter (± 525)	84161 ns/iter (± 596)	2.20
lotr_graph_persistent_window_50_layered/iterate_exploded_edges	4297713 ns/iter (± 87226)	1659940 ns/iter (± 19402)	2.59
lotr_graph_persistent_window_50_layered_materialise/materialize	58397960 ns/iter (± 581105)	5298035 ns/iter (± 147912)	11.02

This comment was automatically generated by workflow using github-action-benchmark.