2.5.20 版本 count(column) count(*) count(1) 数据不一致

[arvin]

【详述】count(column) count(*) count(1) 数据不一致
【背景】无
【业务影响】
【是否存算分离】否
【StarRocks版本】例如：2.5.20
【集群规模】例如：3fe + 8be
【机器信息】
【附件】

count(1) (1.1 KB) count(col) (1.1 KB) count(x) (1.1 KB) countall (1.2 KB)

CREATE TABLE moo (
stat_day date NOT NULL COMMENT “”,
ad_uniqid varchar(32) NOT NULL DEFAULT “0” COMMENT “”,
game_code varchar(8) NOT NULL DEFAULT “” COMMENT " ro_sea ",
app_token varchar(100) NOT NULL DEFAULT “” COMMENT " ",
os varchar(8) NOT NULL DEFAULT “” COMMENT “”,
app_token_name varchar(100) NOT NULL DEFAULT “” COMMENT “”,
country varchar(64) NULL COMMENT “”,
partner varchar(64) NULL COMMENT " ",
campaign varchar(255) NULL COMMENT “”,
adgroup varchar(255) NULL COMMENT “”,
creative varchar(255) NULL COMMENT “”,
campaign_id varchar(255) NULL COMMENT “”,
adgroup_id varchar(255) NULL COMMENT “”,
creative_id varchar(255) NULL COMMENT “”,
account_id varchar(64) NOT NULL DEFAULT “0” COMMENT “”,
image_id varchar(128) NULL COMMENT “”,
video_id varchar(128) NULL COMMENT “”,
updated_at datetime NULL COMMENT “”
) ENGINE=OLAP
PRIMARY KEY(stat_day, ad_uniqid, game_code, app_token, os)
PARTITION BY RANGE(stat_day)
(PARTITION p20230405 VALUES [(“2023-04-05”), (“2023-04-06”)),
…
PARTITION p20240509 VALUES [(“2024-05-09”), (“2024-05-10”)))
DISTRIBUTED BY HASH(ad_uniqid) BUCKETS 32
PROPERTIES (
“replication_num” = “3”,
“dynamic_partition.enable” = “true”,
“dynamic_partition.time_unit” = “DAY”,
“dynamic_partition.time_zone” = “Asia/Shanghai”,
“dynamic_partition.start” = “-2147483648”,
“dynamic_partition.end” = “3”,
“dynamic_partition.prefix” = “p”,
“dynamic_partition.buckets” = “32”,
“dynamic_partition.history_partition_num” = “0”,
“in_memory” = “false”,
“storage_format” = “DEFAULT”,
“enable_persistent_index” = “true”,
“compression” = “LZ4”
);

只要是count(stat_day) 数据就有问题。如果多个count(1),count(stat_day)结果就会正常，单独count数据会多。

[arvin]

@trueeyu

[ranpanfeng]

您可以加我微信吗？ satanson_