Apache Flink 中文用户邮件列表

Flink sql 消费kafka的顺序是怎么样的第二次运行sql的结果和第一次不同

Classic

List

Threaded

5 messages Options

anonnius

Flink sql 消费kafka的顺序是怎么样的第二次运行sql的结果和第一次不同

hi: [求助] 我这里用flink-sql消费kafka数据, 通过窗口做pvuv的计算, 第一次和第二次计算的结果不一致, 不太了解为什么

0> mac本地环境

1> flink 1.11.1

2> kafka 0.10.2.2, topic为message-json, 分区为3, 副本为1

3> 使用的是sql-client.sh 环境

4> 先在sql-cli中创建了iservVisit表

create table iservVisit (

type string comment '时间类型',

uuid string comment '用户uri',

clientTime string comment '10位时间戳',

rowtime as to_timestamp(from_unixtime(cast(substring(coalesce(clientTime, '0'), 1, 10) as bigint))), -- 计算列, 10位时间戳转为timestamp类型

WATERMARK for rowtime as rowtime - INTERVAL '1' MINUTE -- 计算列, 作为watermark

) with (

'connector' = 'kafka-0.10',

'topic' = 'message-json',

'properties.bootstrap.servers' = 'localhost:9092',

'properties.group.id' = 'consumer-rt',

'format' = 'json',

'json.ignore-parse-errors' = 'true',

'scan.startup.mode' = 'earliest-offset'

)

然后在sql-cli执行sql

select

tumble_start(rowtime, interval '2' MINUTE) as wStart,

tumble_end(rowtime, interval '2' MINUTE) as wEnd,

count(1) as pv,

count(distinct uuid) as uv

from iservVisit

group by tumble(rowtime, interval '2' MINUTE)

5> 向kafka生产者依次发送下面的json消息

{"type": "iservVisit", "uuid": "c", "clientTime": "1600391684"}

{"type": "iservVisit", "uuid": "a", "clientTime": "1600391663"}

{"type": "iservVisit", "uuid": "a", "clientTime": "1600391690"}

{"type": "iservVisit", "uuid": "c", "clientTime": "1600391709"}

{"type": "iservVisit", "uuid": "b", "clientTime": "1600391748"}

{"type": "iservVisit", "uuid": "c", "clientTime": "1600391782"}

{"type": "iservVisit", "uuid": "b", "clientTime": "1600391781"}

{"type": "iservVisit", "uuid": "b", "clientTime": "1600391823"}

{"type": "iservVisit", "uuid": "b", "clientTime": "1600391822"}

{"type": "iservVisit", "uuid": "a", "clientTime": "1600391815"}

{"type": "iservVisit", "uuid": "a", "clientTime": "1600391857"}

{"type": "iservVisit", "uuid": "a", "clientTime": "1600391870"}

{"type": "iservVisit", "uuid": "b", "clientTime": "1600391851"}

{"type": "iservVisit", "uuid": "c", "clientTime": "1600391903"}

{"type": "iservVisit", "uuid": "a", "clientTime": "1600391889"}

{"type": "iservVisit", "uuid": "a", "clientTime": "1600391945"}

{"type": "iservVisit", "uuid": "b", "clientTime": "1600391938"}

{"type": "iservVisit", "uuid": "b", "clientTime": "1600391951"}

{"type": "iservVisit", "uuid": "c", "clientTime": "1600391936"}

{"type": "iservVisit", "uuid": "b", "clientTime": "1600391970"}

{"type": "iservVisit", "uuid": "c", "clientTime": "1600392016"}

{"type": "iservVisit", "uuid": "c", "clientTime": "1600391993"}

{"type": "iservVisit", "uuid": "a", "clientTime": "1600392057"}

{"type": "iservVisit", "uuid": "a", "clientTime": "1800392057"}

6> 第一次结果(这里sql-cli的sql一直在运行)

wStart wEnd pv uv

2020-09-18T09:14 2020-09-18T09:16 5 3

2020-09-18T09:16 2020-09-18T09:18 8 3

2020-09-18T09:18 2020-09-18T09:20 8 3

2020-09-18T09:20 2020-09-18T09:22 2 2

7> 第二次结果(退出[Quit]sql-cli中的sql, 在次运行)

wStart wEnd pv uv

2020-09-18T09:14 2020-09-18T09:16 2 2

2020-09-18T09:16 2020-09-18T09:18 2 2

2020-09-18T09:18 2020-09-18T09:20 8 3

2020-09-18T09:20 2020-09-18T09:22 2 2

8> 详细过程以放入附件文件中

flink-sql-kafka.txt (13K) Download Attachment

Evan

回复: Flink sql 消费kafka的顺序是怎么样的第二次运行sql的结果和第一次不同

先占个楼
我按照题主给的文档，一边发送数据，一边执行以下SQL实时查看查询结果
select
tumble_start(rowtime, interval '2' MINUTE) as wStart,
tumble_end(rowtime, interval '2' MINUTE) as wEnd,
count(1) as pv,
count(distinct uuid) as uv
from iservVisit
group by tumble(rowtime, interval '2' MINUTE)
最后得到的结果是这样的：（跟题主不一样）

wStart wEnd pv uv
2020-09-18T09:14 2020-09-18T09:16 2 2
2020-09-18T09:16 2020-09-18T09:18 8 3
2020-09-18T09:18 2020-09-18T09:20 8 3
2020-09-18T09:20 2020-09-18T09:22 2 2

等所有数据都发送完，退出sql-client然后再执行上边的查询语句最后得到的结果：（跟题主是一样的）：
wStart wEnd pv uv
2020-09-18T09:14 2020-09-18T09:16 2 2
2020-09-18T09:16 2020-09-18T09:18 2 2
2020-09-18T09:18 2020-09-18T09:20 8 3
2020-09-18T09:20 2020-09-18T09:22 2 2

发件人： anonnius
发送时间： 2020-09-18 11:24
收件人： user-zh
主题： Flink sql 消费kafka的顺序是怎么样的第二次运行sql的结果和第一次不同
hi: [求助] 我这里用flink-sql消费kafka数据, 通过窗口做pvuv的计算, 第一次和第二次计算的结果不一致, 不太了解为什么
0> mac本地环境
1> flink 1.11.1
2> kafka 0.10.2.2, topic为message-json, 分区为3, 副本为1
3> 使用的是sql-client.sh 环境
4> 先在sql-cli中创建了iservVisit表
create table iservVisit (
type string comment '时间类型',
uuid string comment '用户uri',
clientTime string comment '10位时间戳',
rowtime as to_timestamp(from_unixtime(cast(substring(coalesce(clientTime, '0'), 1, 10) as bigint))), -- 计算列, 10位时间戳转为timestamp类型
WATERMARK for rowtime as rowtime - INTERVAL '1' MINUTE -- 计算列, 作为watermark
) with (
'connector' = 'kafka-0.10',
'topic' = 'message-json',
'properties.bootstrap.servers' = 'localhost:9092',
'properties.group.id' = 'consumer-rt',
'format' = 'json',
'json.ignore-parse-errors' = 'true',
'scan.startup.mode' = 'earliest-offset'
)
然后在sql-cli执行sql
select
tumble_start(rowtime, interval '2' MINUTE) as wStart,
tumble_end(rowtime, interval '2' MINUTE) as wEnd,
count(1) as pv,
count(distinct uuid) as uv
from iservVisit
group by tumble(rowtime, interval '2' MINUTE)
5> 向kafka生产者依次发送下面的json消息
{"type": "iservVisit", "uuid": "c", "clientTime": "1600391684"}
{"type": "iservVisit", "uuid": "a", "clientTime": "1600391663"}
{"type": "iservVisit", "uuid": "a", "clientTime": "1600391690"}
{"type": "iservVisit", "uuid": "c", "clientTime": "1600391709"}
{"type": "iservVisit", "uuid": "b", "clientTime": "1600391748"}
{"type": "iservVisit", "uuid": "c", "clientTime": "1600391782"}
{"type": "iservVisit", "uuid": "b", "clientTime": "1600391781"}
{"type": "iservVisit", "uuid": "b", "clientTime": "1600391823"}
{"type": "iservVisit", "uuid": "b", "clientTime": "1600391822"}
{"type": "iservVisit", "uuid": "a", "clientTime": "1600391815"}
{"type": "iservVisit", "uuid": "a", "clientTime": "1600391857"}
{"type": "iservVisit", "uuid": "a", "clientTime": "1600391870"}
{"type": "iservVisit", "uuid": "b", "clientTime": "1600391851"}
{"type": "iservVisit", "uuid": "c", "clientTime": "1600391903"}
{"type": "iservVisit", "uuid": "a", "clientTime": "1600391889"}
{"type": "iservVisit", "uuid": "a", "clientTime": "1600391945"}
{"type": "iservVisit", "uuid": "b", "clientTime": "1600391938"}
{"type": "iservVisit", "uuid": "b", "clientTime": "1600391951"}
{"type": "iservVisit", "uuid": "c", "clientTime": "1600391936"}
{"type": "iservVisit", "uuid": "b", "clientTime": "1600391970"}
{"type": "iservVisit", "uuid": "c", "clientTime": "1600392016"}
{"type": "iservVisit", "uuid": "c", "clientTime": "1600391993"}
{"type": "iservVisit", "uuid": "a", "clientTime": "1600392057"}
{"type": "iservVisit", "uuid": "a", "clientTime": "1800392057"}
6> 第一次结果(这里sql-cli的sql一直在运行)
wStart wEnd pv uv
2020-09-18T09:14 2020-09-18T09:16 5 3
2020-09-18T09:16 2020-09-18T09:18 8 3
2020-09-18T09:18 2020-09-18T09:20 8 3
2020-09-18T09:20 2020-09-18T09:22 2 2
7> 第二次结果(退出[Quit]sql-cli中的sql, 在次运行)
wStart wEnd pv uv
2020-09-18T09:14 2020-09-18T09:16 2 2
2020-09-18T09:16 2020-09-18T09:18 2 2
2020-09-18T09:18 2020-09-18T09:20 8 3
2020-09-18T09:20 2020-09-18T09:22 2 2
8> 详细过程以放入附件文件中

Evan

回复: Flink sql 消费kafka的顺序是怎么样的第二次运行sql的结果和第一次不同

In reply to this post by anonnius

Hi，问题已经找到了
你的kafka是3个分区，第一次消费你是边发送数据边消费，这时可以认为watermark就是按照你发送数据的顺序生成的，会按照你发送数据的顺序触发计算，所以得到的结果也是你预想的结果。

第二次消费，你的数据不再生产了，这时kafka中的数据可认为是历史数据，你的scan.startup.mode设置的是earliest-offset，这时候flink消费的也是历史数据，因为是3个分区，所以flink会先消费完1个分区内的数据，然后再依次消费另外2个分区的数据，kafka跨分区不能保证有序，所以这时候watermark是按照flink消费数据的顺序生成的，也就是按照如下顺序生成watermark的(此时部分数据会被当做迟到的数据遗弃掉)：

select * from iservVisit
type uuid clientTime rowtime
iservVisit a 1600391663 2020-09-18T09:14:23
iservVisit b 1600391748 2020-09-18T09:15:48
iservVisit b 1600391823 2020-09-18T09:17:03
---此时触发2020-09-18 09:14 - 2020-09-18 09:16
iservVisit a 1600391857 2020-09-18T09:17:37
iservVisit c 1600391903 2020-09-18T09:18:23
iservVisit b 1600391938 2020-09-18T09:18:58
iservVisit b 1600391970 2020-09-18T09:19:30
---此时触发2020-09-18 09:16 - 2020-09-18 09:18
iservVisit a 1600392057 2020-09-18T09:20:57
iservVisit c 1600391684 2020-09-18T09:14:44
iservVisit c 1600391709 2020-09-18T09:15:09
iservVisit b 1600391781 2020-09-18T09:16:21
iservVisit a 1600391815 2020-09-18T09:16:55
iservVisit b 1600391851 2020-09-18T09:17:31
iservVisit a 1600391945 2020-09-18T09:19:05
iservVisit c 1600391936 2020-09-18T09:18:56
iservVisit c 1600391993 2020-09-18T09:19:53
iservVisit a 1600391690 2020-09-18T09:14:50
iservVisit c 1600391782 2020-09-18T09:16:22
iservVisit b 1600391822 2020-09-18T09:17:02
iservVisit a 1600391870 2020-09-18T09:17:50
iservVisit a 1600391889 2020-09-18T09:18:09
iservVisit b 1600391951 2020-09-18T09:19:11
iservVisit c 1600392016 2020-09-18T09:20:16
iservVisit a 1800392057 2027-01-20T04:54:17
---此时触发2020-09-18 09:18 - 2020-09-18 09:20 以及 2020-09-18 09:20 - 2020-09-18 09:22

PS：你可以把你的topic设置成1个分区，这样就可以保证数据整体有序，每次查询得到的结果正确且一样的了。
如果分析的有误，敬请指正！

发件人： anonnius
发送时间： 2020-09-18 11:24
收件人： user-zh
主题： Flink sql 消费kafka的顺序是怎么样的第二次运行sql的结果和第一次不同
hi: [求助] 我这里用flink-sql消费kafka数据, 通过窗口做pvuv的计算, 第一次和第二次计算的结果不一致, 不太了解为什么
0> mac本地环境
1> flink 1.11.1
2> kafka 0.10.2.2, topic为message-json, 分区为3, 副本为1
3> 使用的是sql-client.sh 环境
4> 先在sql-cli中创建了iservVisit表
create table iservVisit (
type string comment '时间类型',
uuid string comment '用户uri',
clientTime string comment '10位时间戳',
rowtime as to_timestamp(from_unixtime(cast(substring(coalesce(clientTime, '0'), 1, 10) as bigint))), -- 计算列, 10位时间戳转为timestamp类型
WATERMARK for rowtime as rowtime - INTERVAL '1' MINUTE -- 计算列, 作为watermark
) with (
'connector' = 'kafka-0.10',
'topic' = 'message-json',
'properties.bootstrap.servers' = 'localhost:9092',
'properties.group.id' = 'consumer-rt',
'format' = 'json',
'json.ignore-parse-errors' = 'true',
'scan.startup.mode' = 'earliest-offset'
)
然后在sql-cli执行sql
select
tumble_start(rowtime, interval '2' MINUTE) as wStart,
tumble_end(rowtime, interval '2' MINUTE) as wEnd,
count(1) as pv,
count(distinct uuid) as uv
from iservVisit
group by tumble(rowtime, interval '2' MINUTE)
5> 向kafka生产者依次发送下面的json消息
{"type": "iservVisit", "uuid": "c", "clientTime": "1600391684"}
{"type": "iservVisit", "uuid": "a", "clientTime": "1600391663"}
{"type": "iservVisit", "uuid": "a", "clientTime": "1600391690"}
{"type": "iservVisit", "uuid": "c", "clientTime": "1600391709"}
{"type": "iservVisit", "uuid": "b", "clientTime": "1600391748"}
{"type": "iservVisit", "uuid": "c", "clientTime": "1600391782"}
{"type": "iservVisit", "uuid": "b", "clientTime": "1600391781"}
{"type": "iservVisit", "uuid": "b", "clientTime": "1600391823"}
{"type": "iservVisit", "uuid": "b", "clientTime": "1600391822"}
{"type": "iservVisit", "uuid": "a", "clientTime": "1600391815"}
{"type": "iservVisit", "uuid": "a", "clientTime": "1600391857"}
{"type": "iservVisit", "uuid": "a", "clientTime": "1600391870"}
{"type": "iservVisit", "uuid": "b", "clientTime": "1600391851"}
{"type": "iservVisit", "uuid": "c", "clientTime": "1600391903"}
{"type": "iservVisit", "uuid": "a", "clientTime": "1600391889"}
{"type": "iservVisit", "uuid": "a", "clientTime": "1600391945"}
{"type": "iservVisit", "uuid": "b", "clientTime": "1600391938"}
{"type": "iservVisit", "uuid": "b", "clientTime": "1600391951"}
{"type": "iservVisit", "uuid": "c", "clientTime": "1600391936"}
{"type": "iservVisit", "uuid": "b", "clientTime": "1600391970"}
{"type": "iservVisit", "uuid": "c", "clientTime": "1600392016"}
{"type": "iservVisit", "uuid": "c", "clientTime": "1600391993"}
{"type": "iservVisit", "uuid": "a", "clientTime": "1600392057"}
{"type": "iservVisit", "uuid": "a", "clientTime": "1800392057"}
6> 第一次结果(这里sql-cli的sql一直在运行)
wStart wEnd pv uv
2020-09-18T09:14 2020-09-18T09:16 5 3
2020-09-18T09:16 2020-09-18T09:18 8 3
2020-09-18T09:18 2020-09-18T09:20 8 3
2020-09-18T09:20 2020-09-18T09:22 2 2
7> 第二次结果(退出[Quit]sql-cli中的sql, 在次运行)
wStart wEnd pv uv
2020-09-18T09:14 2020-09-18T09:16 2 2
2020-09-18T09:16 2020-09-18T09:18 2 2
2020-09-18T09:18 2020-09-18T09:20 8 3
2020-09-18T09:20 2020-09-18T09:22 2 2
8> 详细过程以放入附件文件中

anonnius

Re:回复: Flink sql 消费kafka的顺序是怎么样的第二次运行sql的结果和第一次不同

In reply to this post by anonnius

hi: 感觉你的关注和回复

1> 下面是我的分析过程

1. 第一次是, 先在sql-client.sh 中执行sql

select

tumble_start(rowtime, interval '2' MINUTE) as wStart,

tumble_end(rowtime, interval '2' MINUTE) as wEnd,

count(1) as pv,

count(distinct uuid) as uv

from iservVisit

group by tumble(rowtime, interval '2' MINUTE)

此时, 由于数据是一条一条的通过kafka生产者工具(kafka-console-producer.sh)写入, 并且由kafka-connector会不停的消费数据, 获取的数据是和手动写入的数据的顺序是一样的

2. 第二次是, 退出sql-client.sh后在执行sql

select

tumble_start(rowtime, interval '2' MINUTE) as wStart,

tumble_end(rowtime, interval '2' MINUTE) as wEnd,

count(1) as pv,

count(distinct uuid) as uv

from iservVisit

group by tumble(rowtime, interval '2' MINUTE)

这时由于数据已经写入kafka了, 在由kafka-connector进行消费的时候, 由于topic有3个分区, 消费后获取的消息的顺序和手动通过kafka生产者工具(kafka-console-producer.sh)写入时的顺序

不一致了, 这样rowtime时间靠后的数据可能先被消费, 导致产生了比较大的watermark, 导致后续消费的部分消息被忽略了

3. 通过将建表时 watermark的间隔变大些, 能还原第一次的结果, 这种方式还是考虑中(考虑是否一致有效)

create table iservVisit (

type string comment '时间类型',

uuid string comment '用户uri',

clientTime string comment '10位时间戳',

rowtime as to_timestamp(from_unixtime(cast(substring(coalesce(clientTime, '0'), 1, 10) as bigint))), -- 计算列, 10位时间戳转为timestamp类型

WATERMARK for rowtime as rowtime - INTERVAL '5' MINUTE -- 计算列, 作为watermark, 有1分钟变为5分钟

) with (

'connector' = 'kafka-0.10',

'topic' = 'message-json',

'properties.bootstrap.servers' = 'localhost:9092',

'properties.group.id' = 'consumer-rt',

'format' = 'json',

'json.ignore-parse-errors' = 'true',

'scan.startup.mode' = 'earliest-offset'

)

4. 初步结论是: 如何保证/或通过什么办法, 让每个分区的消费数据的速度保持一致

5. 附件可以通过sublime sql/hql插件查看, 这样显示会清晰点


在 2020-09-18 14:42:42，"[hidden email]" <[hidden email]> 写道：
>先占个楼
>我按照题主给的文档，一边发送数据，一边执行以下SQL实时查看查询结果
>select  
>    tumble_start(rowtime, interval '2' MINUTE) as wStart,
>    tumble_end(rowtime, interval '2' MINUTE) as wEnd,
>    count(1) as pv,
>    count(distinct uuid) as uv 
>from iservVisit
>group by tumble(rowtime, interval '2' MINUTE)
>最后得到的结果是这样的 ：（跟题主不一样）
>
>                 wStart                      wEnd                        pv                        uv
>          2020-09-18T09:14          2020-09-18T09:16                         2                         2
>          2020-09-18T09:16          2020-09-18T09:18                         8                         3
>          2020-09-18T09:18          2020-09-18T09:20                         8                         3
>          2020-09-18T09:20          2020-09-18T09:22                         2                         2
>
>等所有数据都发送完，退出sql-client然后再执行上边的查询语句最后得到的结果：（跟题主是一样的）：
>wStart                                        wEnd                           pv                        uv
>2020-09-18T09:14          2020-09-18T09:16                  2                         2
>2020-09-18T09:16          2020-09-18T09:18                  2                         2
>2020-09-18T09:18          2020-09-18T09:20                  8                         3
>2020-09-18T09:20          2020-09-18T09:22                  2                         2
>
>
>
> 
>发件人： anonnius
>发送时间： 2020-09-18 11:24
>收件人： user-zh
>主题： Flink sql 消费kafka的顺序是怎么样的 第二次运行sql的结果和第一次不同
>hi: [求助] 我这里用flink-sql消费kafka数据, 通过窗口做pvuv的计算, 第一次和第二次计算的结果不一致, 不太了解为什么
>0> mac本地环境
>1> flink 1.11.1
>2> kafka 0.10.2.2, topic为message-json, 分区为3, 副本为1
>3> 使用的是sql-client.sh 环境
>4> 先在sql-cli中创建了iservVisit表
>create table iservVisit (
>    type string comment '时间类型',
>    uuid string comment '用户uri',
>    clientTime string comment '10位时间戳',
>    rowtime as to_timestamp(from_unixtime(cast(substring(coalesce(clientTime, '0'), 1, 10) as bigint))), -- 计算列, 10位时间戳转为timestamp类型
>    WATERMARK for rowtime as rowtime - INTERVAL '1' MINUTE -- 计算列, 作为watermark
>) with (
>    'connector' = 'kafka-0.10',
>    'topic' = 'message-json',
>    'properties.bootstrap.servers' = 'localhost:9092',
>    'properties.group.id' = 'consumer-rt',
>    'format' = 'json',
>    'json.ignore-parse-errors' = 'true',
>    'scan.startup.mode' = 'earliest-offset'
>)
>然后在sql-cli执行sql
>select  
>    tumble_start(rowtime, interval '2' MINUTE) as wStart,
>    tumble_end(rowtime, interval '2' MINUTE) as wEnd,
>    count(1) as pv,
>    count(distinct uuid) as uv 
>from iservVisit
>group by tumble(rowtime, interval '2' MINUTE)
>5> 向kafka生产者依次发送下面的json消息
>{"type": "iservVisit", "uuid": "c", "clientTime": "1600391684"} 
>{"type": "iservVisit", "uuid": "a", "clientTime": "1600391663"} 
>{"type": "iservVisit", "uuid": "a", "clientTime": "1600391690"} 
>{"type": "iservVisit", "uuid": "c", "clientTime": "1600391709"} 
>{"type": "iservVisit", "uuid": "b", "clientTime": "1600391748"} 
>{"type": "iservVisit", "uuid": "c", "clientTime": "1600391782"} 
>{"type": "iservVisit", "uuid": "b", "clientTime": "1600391781"} 
>{"type": "iservVisit", "uuid": "b", "clientTime": "1600391823"} 
>{"type": "iservVisit", "uuid": "b", "clientTime": "1600391822"} 
>{"type": "iservVisit", "uuid": "a", "clientTime": "1600391815"} 
>{"type": "iservVisit", "uuid": "a", "clientTime": "1600391857"} 
>{"type": "iservVisit", "uuid": "a", "clientTime": "1600391870"} 
>{"type": "iservVisit", "uuid": "b", "clientTime": "1600391851"} 
>{"type": "iservVisit", "uuid": "c", "clientTime": "1600391903"} 
>{"type": "iservVisit", "uuid": "a", "clientTime": "1600391889"} 
>{"type": "iservVisit", "uuid": "a", "clientTime": "1600391945"} 
>{"type": "iservVisit", "uuid": "b", "clientTime": "1600391938"} 
>{"type": "iservVisit", "uuid": "b", "clientTime": "1600391951"} 
>{"type": "iservVisit", "uuid": "c", "clientTime": "1600391936"} 
>{"type": "iservVisit", "uuid": "b", "clientTime": "1600391970"} 
>{"type": "iservVisit", "uuid": "c", "clientTime": "1600392016"} 
>{"type": "iservVisit", "uuid": "c", "clientTime": "1600391993"} 
>{"type": "iservVisit", "uuid": "a", "clientTime": "1600392057"} 
>{"type": "iservVisit", "uuid": "a", "clientTime": "1800392057"} 
>6> 第一次结果(这里sql-cli的sql一直在运行)
>    wStart                                      wEnd                        pv                        uv
>2020-09-18T09:14          2020-09-18T09:16                  5                         3
>2020-09-18T09:16          2020-09-18T09:18                  8                         3
>2020-09-18T09:18          2020-09-18T09:20                  8                         3
>2020-09-18T09:20          2020-09-18T09:22                  2                         2
>7> 第二次结果(退出[Quit]sql-cli中的sql, 在次运行)
>wStart                                        wEnd                           pv                        uv
>2020-09-18T09:14          2020-09-18T09:16                  2                         2
>2020-09-18T09:16          2020-09-18T09:18                  2                         2
>2020-09-18T09:18          2020-09-18T09:20                  8                         3
>2020-09-18T09:20          2020-09-18T09:22                  2                         2
>8> 详细过程以放入附件文件中
>
>
>
>
>
>

flink-sql-kafka.txt (15K) Download Attachment

nobleyd

Re: 回复: Flink sql 消费kafka的顺序是怎么样的第二次运行sql的结果和第一次不同

有人解答下，flink sql情况下的watermark生成是否有datastream api中的多分区取小机制呢？

这个问题datastream api是肯定不存在的。
情况1: 如果10个分区，来10个并发即可，然后在后续跟上watermark生成，本身watermark会合并取小。

情况2: 即使是2个并发，每个并发消费5个分区，但只要利用kafkaSouce提供的watermark生成机制也不会有这个问题。

anonnius <[hidden email]> 于2020年9月18日周五下午3:47写道：

> hi: 感觉你的关注和回复
> 1> 下面是我的分析过程
> 1. 第一次是, 先在sql-client.sh 中执行sql
> select
> tumble_start(rowtime, interval '2' MINUTE) as wStart,
> tumble_end(rowtime, interval '2' MINUTE) as wEnd,
> count(1) as pv,
> count(distinct uuid) as uv
> from iservVisit
> group by tumble(rowtime, interval '2' MINUTE)
>
> 此时, 由于数据是一条一条的通过kafka生产者工具(kafka-console-producer.sh)写入,
> 并且由kafka-connector会不停的消费数据, 获取的数据是和手动写入的数据的顺序是一样的
>
> 2. 第二次是, 退出sql-client.sh后在执行sql
> select
> tumble_start(rowtime, interval '2' MINUTE) as wStart,
> tumble_end(rowtime, interval '2' MINUTE) as wEnd,
> count(1) as pv,
> count(distinct uuid) as uv
> from iservVisit
> group by tumble(rowtime, interval '2' MINUTE)
> 这时由于数据已经写入kafka了, 在由kafka-connector进行消费的时候, 由于topic有3个分区, 消费后获取的消息的顺序和
> 手动通过kafka生产者工具(kafka-console-producer.sh)写入时的顺序
> 不一致了, 这样rowtime时间靠后的数据可能先被消费, 导致产生了比较大的watermark, 导致后续消费的部分消息被忽略了
>
> 3. 通过将建表时 watermark的间隔变大些, 能还原第一次的结果, 这种方式还是考虑中(考虑是否一致有效)
> create table iservVisit (
> type string comment '时间类型',
> uuid string comment '用户uri',
> clientTime string comment '10位时间戳',
> rowtime as
> to_timestamp(from_unixtime(cast(substring(coalesce(clientTime, '0'), 1, 10)
> as bigint))), -- 计算列, 10位时间戳转为timestamp类型
> WATERMARK for rowtime as rowtime - INTERVAL '5' MINUTE -- 计算列,
> 作为watermark, 有1分钟变为5分钟
> ) with (
> 'connector' = 'kafka-0.10',
> 'topic' = 'message-json',
> 'properties.bootstrap.servers' = 'localhost:9092',
> 'properties.group.id' = 'consumer-rt',
> 'format' = 'json',
> 'json.ignore-parse-errors' = 'true',
> 'scan.startup.mode' = 'earliest-offset'
> )
> 4. 初步结论是: 如何保证/或通过什么办法, 让每个分区的消费数据的速度保持一致
> 5. 附件可以通过sublime sql/hql插件查看, 这样显示会清晰点
>
>
>
>
>
>
>
> 在 2020-09-18 14:42:42，"[hidden email]" <[hidden email]> 写道：
> >先占个楼
> >我按照题主给的文档，一边发送数据，一边执行以下SQL实时查看查询结果
> >select
> > tumble_start(rowtime, interval '2' MINUTE) as wStart,
> > tumble_end(rowtime, interval '2' MINUTE) as wEnd,
> > count(1) as pv,
> > count(distinct uuid) as uv
> >from iservVisit
> >group by tumble(rowtime, interval '2' MINUTE)
> >最后得到的结果是这样的：（跟题主不一样）
> >
> > wStart wEnd pv uv
> > 2020-09-18T09:14 2020-09-18T09:16 2 2
> > 2020-09-18T09:16 2020-09-18T09:18 8 3
> > 2020-09-18T09:18 2020-09-18T09:20 8 3
> > 2020-09-18T09:20 2020-09-18T09:22 2 2
> >
> >等所有数据都发送完，退出sql-client然后再执行上边的查询语句最后得到的结果：（跟题主是一样的）：
> >wStart wEnd pv uv
> >2020-09-18T09:14 2020-09-18T09:16 2 2
> >2020-09-18T09:16 2020-09-18T09:18 2 2
> >2020-09-18T09:18 2020-09-18T09:20 8 3
> >2020-09-18T09:20 2020-09-18T09:22 2 2
> >
> >
> >
> >
> >发件人： anonnius
> >发送时间： 2020-09-18 11:24
> >收件人： user-zh
> >主题： Flink sql 消费kafka的顺序是怎么样的第二次运行sql的结果和第一次不同
> >hi: [求助] 我这里用flink-sql消费kafka数据, 通过窗口做pvuv的计算, 第一次和第二次计算的结果不一致, 不太了解为什么
> >0> mac本地环境
> >1> flink 1.11.1
> >2> kafka 0.10.2.2, topic为message-json, 分区为3, 副本为1
> >3> 使用的是sql-client.sh 环境
> >4> 先在sql-cli中创建了iservVisit表
> >create table iservVisit (
> > type string comment '时间类型',
> > uuid string comment '用户uri',
> > clientTime string comment '10位时间戳',
> > rowtime as to_timestamp(from_unixtime(cast(substring(coalesce(clientTime, '0'), 1, 10) as bigint))), -- 计算列, 10位时间戳转为timestamp类型
> > WATERMARK for rowtime as rowtime - INTERVAL '1' MINUTE -- 计算列, 作为watermark
> >) with (
> > 'connector' = 'kafka-0.10',
> > 'topic' = 'message-json',
> > 'properties.bootstrap.servers' = 'localhost:9092',
> > 'properties.group.id' = 'consumer-rt',
> > 'format' = 'json',
> > 'json.ignore-parse-errors' = 'true',
> > 'scan.startup.mode' = 'earliest-offset'
> >)
> >然后在sql-cli执行sql
> >select
> > tumble_start(rowtime, interval '2' MINUTE) as wStart,
> > tumble_end(rowtime, interval '2' MINUTE) as wEnd,
> > count(1) as pv,
> > count(distinct uuid) as uv
> >from iservVisit
> >group by tumble(rowtime, interval '2' MINUTE)
> >5> 向kafka生产者依次发送下面的json消息
> >{"type": "iservVisit", "uuid": "c", "clientTime": "1600391684"}
> >{"type": "iservVisit", "uuid": "a", "clientTime": "1600391663"}
> >{"type": "iservVisit", "uuid": "a", "clientTime": "1600391690"}
> >{"type": "iservVisit", "uuid": "c", "clientTime": "1600391709"}
> >{"type": "iservVisit", "uuid": "b", "clientTime": "1600391748"}
> >{"type": "iservVisit", "uuid": "c", "clientTime": "1600391782"}
> >{"type": "iservVisit", "uuid": "b", "clientTime": "1600391781"}
> >{"type": "iservVisit", "uuid": "b", "clientTime": "1600391823"}
> >{"type": "iservVisit", "uuid": "b", "clientTime": "1600391822"}
> >{"type": "iservVisit", "uuid": "a", "clientTime": "1600391815"}
> >{"type": "iservVisit", "uuid": "a", "clientTime": "1600391857"}
> >{"type": "iservVisit", "uuid": "a", "clientTime": "1600391870"}
> >{"type": "iservVisit", "uuid": "b", "clientTime": "1600391851"}
> >{"type": "iservVisit", "uuid": "c", "clientTime": "1600391903"}
> >{"type": "iservVisit", "uuid": "a", "clientTime": "1600391889"}
> >{"type": "iservVisit", "uuid": "a", "clientTime": "1600391945"}
> >{"type": "iservVisit", "uuid": "b", "clientTime": "1600391938"}
> >{"type": "iservVisit", "uuid": "b", "clientTime": "1600391951"}
> >{"type": "iservVisit", "uuid": "c", "clientTime": "1600391936"}
> >{"type": "iservVisit", "uuid": "b", "clientTime": "1600391970"}
> >{"type": "iservVisit", "uuid": "c", "clientTime": "1600392016"}
> >{"type": "iservVisit", "uuid": "c", "clientTime": "1600391993"}
> >{"type": "iservVisit", "uuid": "a", "clientTime": "1600392057"}
> >{"type": "iservVisit", "uuid": "a", "clientTime": "1800392057"}
> >6> 第一次结果(这里sql-cli的sql一直在运行)
> > wStart wEnd pv uv
> >2020-09-18T09:14 2020-09-18T09:16 5 3
> >2020-09-18T09:16 2020-09-18T09:18 8 3
> >2020-09-18T09:18 2020-09-18T09:20 8 3
> >2020-09-18T09:20 2020-09-18T09:22 2 2
> >7> 第二次结果(退出[Quit]sql-cli中的sql, 在次运行)
> >wStart wEnd pv uv
> >2020-09-18T09:14 2020-09-18T09:16 2 2
> >2020-09-18T09:16 2020-09-18T09:18 2 2
> >2020-09-18T09:18 2020-09-18T09:20 8 3
> >2020-09-18T09:20 2020-09-18T09:22 2 2
> >8> 详细过程以放入附件文件中
> >
> >
> >
> >
> >
> >
>
>
>
>
>
>
>
>

Flink sql 消费kafka的顺序是怎么样的 第二次运行sql的结果和第一次不同

Flink sql 消费kafka的顺序是怎么样的 第二次运行sql的结果和第一次不同

回复: Flink sql 消费kafka的顺序是怎么样的 第二次运行sql的结果和第一次不同

回复: Flink sql 消费kafka的顺序是怎么样的 第二次运行sql的结果和第一次不同

Re:回复: Flink sql 消费kafka的顺序是怎么样的 第二次运行sql的结果和第一次不同

Re: 回复: Flink sql 消费kafka的顺序是怎么样的 第二次运行sql的结果和第一次不同

Flink sql 消费kafka的顺序是怎么样的第二次运行sql的结果和第一次不同

Flink sql 消费kafka的顺序是怎么样的第二次运行sql的结果和第一次不同

回复: Flink sql 消费kafka的顺序是怎么样的第二次运行sql的结果和第一次不同

回复: Flink sql 消费kafka的顺序是怎么样的第二次运行sql的结果和第一次不同

Re:回复: Flink sql 消费kafka的顺序是怎么样的第二次运行sql的结果和第一次不同

Re: 回复: Flink sql 消费kafka的顺序是怎么样的第二次运行sql的结果和第一次不同