Do you mean extract data from the fact table? Define a model based on this
table, and dimensions & this measure, then create an aggregation group with
these dimensions and measure, build to get the data you preferred. Or do
you mean the metadata file? Just query the database or dump metadata to
??
----
??: "marc nicole"
the dataset (csv) is in this form
att1
att2
count_measure
str1
str2
int1
...
How to extract from it fact table (the count_measure column) and the
dimensions ? in the datasource / model definition in Kylin?
Le mer. 2 août 2023 à 13:32, marc nicole a écrit :
> in the model the measure
in the model the measure definition is as follows (I don't understand
chinese)
{
"name": "COUNT1",
"function": {
"expression": "COUNT",
"parameter": {
"type": "column",
"value": "attributeName.COUNT1"
},
"returntype": "bigint"
Hey, to unsubscribe user mailling list, please follow these steps:
First, you need to send any text to user-unsubscr...@kylin.apache.org .
And you will receive a confirm email with title "confirm unsubscribe from
user@kylin.apache.org " in several minutes.
Then, you have to reply (with any
退订
你好，由于工作原因退订所有相关的邮件。
谢谢！
----
??: "??"
I guessed Shaofeng has answered your question in wechat group. Following is his
answer:
把维度表中的字段，做成“normal”维度，这样这些字段的数据，就会被持久化到cube中；构建新的segment的时候，也不会改变过去构建的结果了
如果不是“normal”维度，就会存在于维度表的快照中，所以维度表快照被刷新后数据也就变了
请问你为什么不考虑使用一般维度呢（因为设置为一般维度后，就不会生成维度表快照了）？
--
Best wishes to you !
From
问题就是，维度数据每构建一次就需要变动一次，导致了前面构建的也用了最新的维度数据，但是构建时间跨天就不会有这种问题，还是维度快照的更新机制问题，怎么能解决维度数据变了，前面已经构建过的关联的维表还是当时的那种状态呢？
| |
李甜彪
|
|
ltb1...@163.com
|
Replied Message
| From | P.F. ZHAN |
| Date | 8/2/2023 15:34 |
| To | |
| Subject | Re: measure column showing 1 as values instead of the actual
这种你不是应该将需要查询的列设置成维度，然后预计算存储到cube么？cube的维度数据，如果不做刷新，那么就不会变化。
On Wed, Aug 2, 2023 at 11:35 李甜彪 wrote:
> 大神，能帮我解决一下我碰到的问题吗？维度表的快照怎么能不让取最新的，
> 在kylin4.0.3的使用过程中，在同一天构建不同日期分区数据时，维度表需要切换，但是当天的最后一次构建会导致今天构建的其他天数据也使用了最后一次所关联的维度表，现在的想法是每次构建关联维度表的数据在变，但是也要让当天已经构建过的数据不发生变化，有什么方法可以实现。
> 李甜彪
>
10 matches
Mail list logo
