请帮我判断下这是不是存在重复时间主键

开源版问题专区
1

【TDengine 使用环境】
预生产环境

【TDengine 版本】

3.3.8.8

【操作系统以及版本】

Ubuntu

【部署方式】

容器

【集群节点数】

单节点

【集群副本数】

单副本

【描述业务影响】

从A表通过Stream汇总数据到B表时，发现时间戳出现重复

【问题复现路径/shan】

之前Stream出过问题报param error，自动undeploy后，删除旧stream重新建stream启动，今天例行巡检发现时间重复的数据

【遇到的问题：问题现象及影响】

通过stream从A表按秒汇总数据到B表中，理论上时间戳不应该出现重复

【资源配置】

4核心 8G A表和B表都是不到300个子表

【报错完整截图】（不要大段的粘贴报错代码，论坛直接看报错代码不直观）


CREATE STABLE IF NOT EXISTS mod_emo_detect_log(
        ts TIMESTAMP,
        con_id INT COMPOSITE KEY,
        emo INT,
        confi FLOAT,
        weight FLOAT
    ) TAGS (
        user_id INT
    )
    KEEP 90d;


CREATE STABLE IF NOT EXISTS mod_emo_detect_data(
        ts TIMESTAMP,
        con_id INT COMPOSITE KEY,
        emo INT,
        confi FLOAT,
        detect_count INT,
        weight FLOAT
    ) TAGS (
        user_id INT
    )
    KEEP 36500d;



taos> select ts,con_id,emo,weight,confi,user_id from mod_emo_detect_log order by ts desc limit 10;
           ts            |  con_id  |   emo   |   weight    | confi      | user_id   |
======================================================================================
 2026-05-05 14:28:44.803 |       15 |       0 |         0.2 |   0.857427 |       309 |  XXX
 2026-05-05 14:28:39.546 |       15 |       0 |         0.2 |   0.815098 |       381 |
 2026-05-05 14:28:39.546 |       15 |       0 |         0.2 |   0.844953 |       377 |
 2026-05-05 14:28:39.546 |       15 |       0 |         0.2 |   0.806491 |       312 |
 2026-05-05 14:28:39.546 |       15 |       0 |         0.2 |   0.699967 |       295 |
 2026-05-05 14:28:39.546 |       15 |       0 |         0.2 |   0.704526 |       300 |
 2026-05-05 14:28:39.546 |       15 |       0 |         0.2 |   0.567211 |       398 |  XXXXXX
 2026-05-05 14:28:38.619 |       15 |       0 |         0.2 |   0.689002 |       398 |
 2026-05-05 14:28:38.619 |       15 |       0 |         0.2 |   0.707646 |       300 |
 2026-05-05 14:28:38.619 |       15 |       0 |         0.2 |   0.677549 |       295 |
Query OK, 10 row(s) in set (0.013765s)


taos> select ts,con_id,emo,weight,confi,detect_count,user_id from mod_emo_detect_data  order by ts desc limit 12;
           ts            |  con_id  |   emo   | weight     | confi      | detect_count | user_id   |
====================================================================================================
 2026-05-05 14:28:44.000 |       15 |       0 |        0.2 |   0.857427 |            1 |       309 |  XXX 子表B309
 2026-05-05 14:28:44.000 |       15 |       0 |        0.2 |   0.857427 |            1 |       309 |  XXX 子表B309
 2026-05-05 14:28:39.000 |       15 |       0 |        0.2 |   0.815098 |            1 |       381 |
 2026-05-05 14:28:39.000 |       15 |       0 |        0.2 |   0.815098 |            1 |       381 |
 2026-05-05 14:28:39.000 |       15 |       0 |        0.2 |   0.704526 |            1 |       300 |
 2026-05-05 14:28:39.000 |       15 |       0 |        0.2 |   0.844953 |            1 |       377 |
 2026-05-05 14:28:39.000 |       15 |       0 |        0.2 |   0.806491 |            1 |       312 |
 2026-05-05 14:28:39.000 |       15 |       0 |        0.2 |   0.699967 |            1 |       295 |
 2026-05-05 14:28:39.000 |       15 |       0 |        0.2 |   0.567211 |            1 |       398 |  XXXXXX 子表B398
 2026-05-05 14:28:39.000 |       15 |       0 |        0.2 |   0.567211 |            1 |       398 |  XXXXXX 子表B398
 2026-05-05 14:28:38.000 |       15 |       0 |        0.2 |   0.689002 |            1 |       398 |
 2026-05-05 14:28:38.000 |       15 |       0 |        0.2 |   0.677549 |            1 |       295 |
Query OK, 12 row(s) in set (0.046969s)

标记的XXX和XXXXXX为存疑相关数据，且mod_emo_detect_data中重复数据确认是在同一张子表中，标记XXX的都在子表B309，标记XXXXXX的都在子表B398