Filter on timestamptz iceberg data returns empty rows (when server timezone is set) #1299
Description
Describe the bug
ClickHouse server returns empty rows(Iceberg rest catalog) when the server is set to a different timezone. There was an issue with ice Altinity/ice#85, where the min/max partition values were wrong.
:) set iceberg_timezone_for_timestamptz ='America/Chicago';
SET iceberg_timezone_for_timestamptz = 'America/Chicago'
Query id: 27156193-fa6b-4dd0-b9e7-7c90d49e085b
Ok.
0 rows in set. Elapsed: 0.002 sec.
:) select count(*), min(_time), max(_time) from `test.test4`;
SELECT
count(*),
min(_time),
max(_time)
FROM `test.test4`
Query id: 5f2bd8c8-01ca-4323-8fd8-97641dd32afe
┌─count()─┬─────────────────min(_time)─┬─────────────────max(_time)─┐
1. │ 86400 │ 2025-12-16 00:00:00.000000 │ 2025-12-16 23:59:59.000000 │
└─────────┴────────────────────────────┴────────────────────────────┘
1 row in set. Elapsed: 0.010 sec. Processed 86.40 thousand rows, 273.82 KB (8.96 million rows/s., 28.39 MB/s.)
Peak memory usage: 822.20 KiB.
:) select count(*), min(_time), max(_time) from `test.test4` where _time >= '2025-12-16 18:05:00' and _time < '2025-12-16 18:10:00' settings use_iceberg_partition_pruning = 1;
SELECT
count(*),
min(_time),
max(_time)
FROM `test.test4`
WHERE (_time >= '2025-12-16 18:05:00') AND (_time < '2025-12-16 18:10:00')
SETTINGS use_iceberg_partition_pruning = 1
Query id: 7296a2d9-e56f-400c-afa0-8ce597e07039
┌─count()─┬─────────────────min(_time)─┬─────────────────max(_time)─┐
1. │ 0 │ 1969-12-31 18:00:00.000000 │ 1969-12-31 18:00:00.000000 │
└─────────┴────────────────────────────┴────────────────────────────┘
1 row in set. Elapsed: 0.009 sec.
^^^^ BUG expected 300 got 0
To Reproduce
Steps to reproduce the behavior:
- Run clickhouse server in a different timezone, example:
America/Chicago - Use
iceto insert iceberg data (partitioned by day)
clickhouse local --query "select toDateTime(toDate(now()) + interval number second) as _time, 1 as value from numbers(86400) format Parquet" > test.parquet
./ice insert test.test4 -p test.parquet --partition='[{"column":"_time", "transform": "day"}]'
./ice insert test.test5 -p test.parquet
- SET
set iceberg_timezone_for_timestamptz ='America/Chicago';
Expected behavior
- ClickHouse server should be able to retrieve the partition information from iceberg rest catalog and using the partition information, it should be able to filter the results.
Key information
- Not reproducible in non-antalya build and Spark
dfa6d88178f2 :) select count(*), min(_time), max(_time) from `test.test4` where _time >= '2025-12-19 18:05:00' and _time < '2025-12-19 18:10:00' settings use_iceberg_partition_pruning = 1;
SELECT
count(*),
min(_time),
max(_time)
FROM `test.test4`
WHERE (_time >= '2025-12-19 18:05:00') AND (_time < '2025-12-19 18:10:00')
SETTINGS use_iceberg_partition_pruning = 1
Query id: 4ca84779-c106-40ad-b19a-e0b8d1308544
┌─count()─┬─────────────────min(_time)─┬─────────────────max(_time)─┐
1. │ 300 │ 2025-12-19 18:05:00.000000 │ 2025-12-19 18:09:59.000000 │
└─────────┴────────────────────────────┴────────────────────────────┘
1 row in set. Elapsed: 0.025 sec. Processed 86.40 thousand rows, 273.89 KB (3.46 million rows/s., 10.97 MB/s.)
Peak memory usage: 1.39 MiB.
dfa6d88178f2 :) select timezone();
SELECT timezone()
Query id: 48cf05e8-6bbf-4684-a1be-c7f6230477f2
┌─timezone()──────┐
1. │ America/Chicago │
└─────────────────┘
1 row in set. Elapsed: 0.003 sec.
dfa6d88178f2 :) select version();
SELECT version()
Query id: 701e5317-a481-4fe2-9d28-11d73b18749e
┌─version()──┐
1. │ 25.8.14.17 │
└────────────┘
1 row in set. Elapsed: 0.002 sec.