1

使用下面的 DDL 和配置文件 yaml，我使用cassandra-stress. 我得到的列结果amount与status预期不符。随机值似乎是每个分区绘制一次，而不是每一行。

例如，如果cassandra-stress生成具有相同business_date（即一个分区）的5 行，amount并且status值重复 5 次，则“下一个”随机值在business_date更改时出现。amount我怎样才能做到这一点，以便我status每行都获得新的平局？

示例输出，请注意仅在第一列更改后最后两列更改值。

2018-09-26,y~8.>6MZ,00000000-0004-0a3c-0000-000000040a3c,5.133114565746717E10,3PR|I{3B
2018-09-26,y~8.>6MZ,00000000-004c-4e7e-0000-0000004c4e7e,5.133114565746717E10,3PR|I{3B
2018-09-26,y~8.>6MZ,00000000-003d-b97f-0000-0000003db97f,5.133114565746717E10,3PR|I{3B
2018-09-26,y~8.>6MZ,00000000-004f-db3f-0000-0000004fdb3f,5.133114565746717E10,3PR|I{3B
2018-09-26,y~8.>6MZ,00000000-008c-f0ea-0000-0000008cf0ea,5.133114565746717E10,3PR|I{3B
2018-10-14,Y ?R|    |u,00000000-002b-5707-0000-0000002b5707,6.698617679577381E10,,fkb[cU~N!
.
.
.

表结构：

CREATE TABLE IF NOT EXISTS record (
business_date date,
region text,
id uuid,
status text,
amount double,
PRIMARY KEY (business_date, region, id)
);

简介 YAML：

keyspace: dev
table: record
columnspec:
 - name: business_date
   population: uniform(17800..17845)
 - name: region
   size: fixed(10)
   population: seq(10..16)
   cluster: fixed(7)
 - name: id
   size: fixed(32)
   population: seq(1..10M)
   cluster: fixed(5)
 - name: status
   size: fixed(10)
   population: uniform(1000..1010)
 - name: amount
   population: uniform(500000..10M)
insert:
   partitions: fixed(1)
   select: fixed(1)/35
queries:
   selectall:
    cql: select * from record where business_date = ? and region = ?
    fields : samerow
4

0 回答 0