Improve iceberg hive statistics documentation

[ZacBlanco]

Description

• Documented the configuration properties for merging statistics from the hive metastore.
• Added a single integration test to ensure stats are always stored when ANALYZE is run, regardless of whether the iceberg.hive-statistics-merge-strategy is set

Motivation and Context

Documentation improvements

Impact

N/A

Test Plan

N/A

Contributor checklist

• Please make sure your submission complies with our development, formatting, commit message, and attribution guidelines.
• PR description addresses the issue accurately and concisely. If the change is non-trivial, a GitHub Issue is referenced.
• Documented new properties (with its default value), SQL syntax, functions, or other functionality.
• If release notes are required, they follow the release notes guidelines.
• Adequate tests were added if applicable.
• CI passed.

Release Notes

== NO RELEASE NOTE ==

[github-actions]

Codenotify: Notifying subscribers in CODENOTIFY files for diff 29cc196...dcf052d.

Notify	File(s)
@steveburnett	presto-docs/src/main/sphinx/connector/iceberg.rst

[imjalpreet]

@ZacBlanco This was already documented last week, maybe we can edit that if we want to modify the explanation: https://github.com/prestodb/presto/blame/29cc1963bb6f1f268e4a46044385616e715e2214/presto-docs/src/main/sphinx/connector/iceberg.rst#L228

[ZacBlanco]

Ah thanks. I think my fork wasn't updated. I'll modify this PR

[ethanyzhang]

Can you add a case for a partitioned table?

[ethanyzhang]

This is my test:

❯ presto-cli \
--catalog iceberg \
--schema tpcds_sf1000_parquet_varchar_iceberg_part \
--server https://***************
presto:tpcds_sf1000_parquet_varchar_iceberg_part> set session iceberg.hive_statistics_merge_strategy='USE_NULLS_FRACTION_AND_NDV';
SET SESSION
presto:tpcds_sf1000_parquet_varchar_iceberg_part> show stats for inventory;
     column_name      |   data_size   | distinct_values_count |   nulls_fraction    | row_count | low_value | high_val>
----------------------+---------------+-----------------------+---------------------+-----------+-----------+--------->
 inv_item_sk          | 2.271521964E9 | NULL                  |                 0.0 | NULL      | 1         | 300000  >
 inv_warehouse_sk     |  1.93783029E8 | NULL                  |                 0.0 | NULL      | 1         | 20      >
 inv_quantity_on_hand |     3318960.0 | NULL                  | 0.05000479948914432 | NULL      | 0         | 1000    >
 inv_date_sk          |        6932.0 | NULL                  |                 0.0 | NULL      | 2450815   | 2452635 >
 NULL                 | NULL          | NULL                  | NULL                |    7.83E8 | NULL      | NULL    >
(5 rows)

Query 20240312_081321_00012_65hxe, FINISHED, 1 node
Splits: 1 total, 1 done (100.00%)
[Latency: client-side: 0:04, server-side: 0:04] [0 rows, 0B] [0 rows/s, 0B/s]

presto:tpcds_sf1000_parquet_varchar_iceberg_part> analyze inventory;
ANALYZE: 783000000 rows

Query 20240312_081333_00013_65hxe, FINISHED, 9 nodes
Splits: 295 total, 295 done (100.00%)
[Latency: client-side: 0:20, server-side: 0:19] [783M rows, 2.3GB] [40.4M rows/s, 122MB/s]

presto:tpcds_sf1000_parquet_varchar_iceberg_part> show stats for inventory;
     column_name      |   data_size   | distinct_values_count |   nulls_fraction    | row_count | low_value | high_val>
----------------------+---------------+-----------------------+---------------------+-----------+-----------+--------->
 inv_item_sk          | 2.271521964E9 | NULL                  |                 0.0 | NULL      | 1         | 300000  >
 inv_warehouse_sk     |  1.93783029E8 | NULL                  |                 0.0 | NULL      | 1         | 20      >
 inv_quantity_on_hand |     3318960.0 | NULL                  | 0.05000479948914432 | NULL      | 0         | 1000    >
 inv_date_sk          |        6932.0 | NULL                  |                 0.0 | NULL      | 2450815   | 2452635 >
 NULL                 | NULL          | NULL                  | NULL                |    7.83E8 | NULL      | NULL    >
(5 rows)

Query 20240312_081357_00014_65hxe, FINISHED, 1 node
Splits: 1 total, 1 done (100.00%)
[Latency: client-side: 452ms, server-side: 347ms] [0 rows, 0B] [0 rows/s, 0B/s]

presto:tpcds_sf1000_parquet_varchar_iceberg_part> show create table inventory;
                                    Create Table
------------------------------------------------------------------------------------
 CREATE TABLE iceberg.tpcds_sf1000_parquet_varchar_iceberg_part.inventory (
    "inv_item_sk" integer,
    "inv_warehouse_sk" integer,
    "inv_quantity_on_hand" integer,
    "inv_date_sk" integer
 )
 WITH (
    delete_mode = 'copy-on-write',
    format = 'PARQUET',
    format_version = '2',
    location = 's3a://presto-workload/tpcds-sf1000-parquet-iceberg-part/inventory',
    partitioning = ARRAY['inv_date_sk']
 )
(1 row)

Query 20240312_081417_00015_65hxe, FINISHED, 1 node
Splits: 1 total, 1 done (100.00%)
[Latency: client-side: 209ms, server-side: 104ms] [0 rows, 0B] [0 rows/s, 0B/s]

presto:tpcds_sf1000_parquet_varchar_iceberg_part>

[ZacBlanco]

I realized we don't return NDV stats for partitioned tables. See #22168

[ZacBlanco]

Closing this. I think the current documentation is fine for now