参考:

Phoenix与HBase集成进行数据分析

HBase查询速度慢原因排查

操作1,执行查询,如下:

: jdbc:phoenix:node3::/hbase> SELECT * FROM ASSET_RECORD WHERE ASSET_ID='设345-1149640126759047168';
+-----------------------------------+---------------------------+-------------------+---------------+-----------------------------------+--------+
|                ID                 |         ASSET_ID          | MANAGEMENT_TABLE  | INTRODUCTION  |           MANAGEMENT_ID           |        |
+-----------------------------------+---------------------------+-------------------+---------------+-----------------------------------+--------+
| 0292ebbfdf3e4d97a6e9fc930ed126d4  | 设345-  | ASSET_SEAL        |               | dd9ff0fc0ad4486bb0812e78fa53ce0e  | - |
+-----------------------------------+---------------------------+-------------------+---------------+-----------------------------------+--------+
 row selected (0.081 seconds)

操作2,重复以上查询,如下:

: jdbc:phoenix:node3::/hbase> SELECT * FROM ASSET_RECORD WHERE ASSET_ID='设345-1149640126759047168';
+-----------------------------------+---------------------------+-------------------+---------------+-----------------------------------+--------+
|                ID                 |         ASSET_ID          | MANAGEMENT_TABLE  | INTRODUCTION  |           MANAGEMENT_ID           |        |
+-----------------------------------+---------------------------+-------------------+---------------+-----------------------------------+--------+
| 0292ebbfdf3e4d97a6e9fc930ed126d4  | 设345-  | ASSET_SEAL        |               | dd9ff0fc0ad4486bb0812e78fa53ce0e  | - |
+-----------------------------------+---------------------------+-------------------+---------------+-----------------------------------+--------+
 row selected (0.077 seconds)

操作3,使用explain重复以上查询,如下:

: jdbc:phoenix:node3::/hbase> explain SELECT * FROM ASSET_RECORD WHERE ASSET_ID='设345-1149640126759047168';
+----------------------------------------------------------------------------------------------------+-----------------+----------------+--------+
|                                                PLAN                                                | EST_BYTES_READ  | EST_ROWS_READ  |  EST_I |
+----------------------------------------------------------------------------------------------------+-----------------+----------------+--------+
| CLIENT -CHUNK  ROWS  BYTES PARALLEL -WAY ROUND ROBIN FULL SCAN OVER ASSET_RECORD  |        |          |  |
|     SERVER FILTER BY ASSET_ID =        |          |  |
+----------------------------------------------------------------------------------------------------+-----------------+----------------+--------+
 rows selected (0.015 seconds)

操作4,在表上建索引,如下:

: jdbc:phoenix:node3::/hbase> create index IDX_ASSET_RECORD on ASSET_RECORD(ASSET_ID,MANAGEMENT_TABLE);
, rows affected (6.25 seconds)

操作5,强制使用索引执行查询,如下:

: jdbc:phoenix:node3::/hbase> SELECT /*+ INDEX(ASSET_RECORD IDX_ASSET_RECORD)*/ * FROM ASSET_RECORD WHERE ASSET_ID='设345-1149640126759047168;
+-----------------------------------+---------------------------+-------------------+---------------+-----------------------------------+--------+
|                ID                 |         ASSET_ID          | MANAGEMENT_TABLE  | INTRODUCTION  |           MANAGEMENT_ID           |        |
+-----------------------------------+---------------------------+-------------------+---------------+-----------------------------------+--------+
| 0292ebbfdf3e4d97a6e9fc930ed126d4  | 设345-  | ASSET_SEAL        |               | dd9ff0fc0ad4486bb0812e78fa53ce0e  | - |
+-----------------------------------+---------------------------+-------------------+---------------+-----------------------------------+--------+
 row selected (0.058 seconds)

操作6,强制使用索引执行查询,如下:

: jdbc:phoenix:node3::/hbase> SELECT /*+ INDEX(ASSET_RECORD IDX_ASSET_RECORD)*/ * FROM ASSET_RECORD WHERE ASSET_ID='设345-1149640126759047168';
+-----------------------------------+---------------------------+-------------------+---------------+-----------------------------------+--------+
|                ID                 |         ASSET_ID          | MANAGEMENT_TABLE  | INTRODUCTION  |           MANAGEMENT_ID           |        |
+-----------------------------------+---------------------------+-------------------+---------------+-----------------------------------+--------+
| 0292ebbfdf3e4d97a6e9fc930ed126d4  | 设345-  | ASSET_SEAL        |               | dd9ff0fc0ad4486bb0812e78fa53ce0e  | - |
+-----------------------------------+---------------------------+-------------------+---------------+-----------------------------------+--------+
 row selected (0.033 seconds)

操作7,使用explain强制使用索引执行查询,如下:

: jdbc:phoenix:node3::/hbase> explain SELECT /*+ INDEX(ASSET_RECORD IDX_ASSET_RECORD)*/ * FROM ASSET_RECORD WHERE ASSET_ID='设345-114964012679047168';
+------------------------------------------------------------------------------------------------------------------+-----------------+-----------+
|                                                       PLAN                                                       | EST_BYTES_READ  | EST_ROWS_ |
+------------------------------------------------------------------------------------------------------------------+-----------------+-----------+
| CLIENT -CHUNK  ROWS  BYTES PARALLEL -WAY ROUND ROBIN FULL SCAN OVER ASSET_RECORD                | null            | null      |
|     SKIP-SCAN-JOIN TABLE                                                                                        | null            | null      |
|         CLIENT -CHUNK PARALLEL -WAY ROUND ROBIN RANGE SCAN OVER IDX_ASSET_RECORD ['设345-1149640126759047168']  | null            | null      |
|             SERVER FILTER BY FIRST KEY ONLY                                                                      | null            | null      |
|     DYNAMIC SERVER FILTER BY .$)                                                      | null            | null      |
+------------------------------------------------------------------------------------------------------------------+-----------------+-----------+
 rows selected (0.045 seconds)

操作8,删除索引,如下:

: jdbc:phoenix:node3::/hbase> drop index IDX_ASSET_RECORD on ASSET_RECORD;
No rows affected (3.688 seconds)

计算操作1和操作2的平均执行时间,建索引后,计算操作5和操作6的平均执行时间,经比较发现使用索引确实提高了查询的速度。

Phoenix具有索引同步更新机制,增删改一条或多条数据以后,索引会自动更新;但是,如果原来的表增加了字段,那就需要更新建在表上的索引。

表的属性越多,条目越多,建索引节约的时间越多,如下是82个属性和195821条记录的表:

: jdbc:phoenix:node3::/hbase> SELECT COUNT(*) FROM ASSET_NORMAL;
+-----------+
| COUNT()  |
+-----------+
|     |
+-----------+
 row selected (4.54 seconds)
: jdbc:phoenix:node3::/hbase> create index IDX_ASSET_NORMAL on ASSET_NORMAL(ASSET_ID,ASSET_NAME,USER_ID);
, rows affected (8.887 seconds)
: jdbc:phoenix:node3::/hbase> SELECT /*+ INDEX(ASSET_NORMAL IDX_ASSET_NORMAL)*/ * FROM ASSET_NORMAL WHERE ASSET_ID='仪1-1151470269278326784';
+-----------------------------------+-------------------------+-------------+------------------------+--------------------------+----------------+
|                ID                 |        ASSET_ID         | ASSET_NAME  | ASSET_FIRST_DEGREE_ID  | ASSET_FIRST_DEGREE_NAME  | ASSET_SECOND_D |
+-----------------------------------+-------------------------+-------------+------------------------+--------------------------+----------------+
| 002e028151e24b07a21e0a0e9ce7f74c  | 仪1-  | 测量仪器        |                 | 仪表                       |      |
+-----------------------------------+-------------------------+-------------+------------------------+--------------------------+----------------+
 row selected (0.209 seconds)
: jdbc:phoenix:node3::/hbase> SELECT * FROM ASSET_NORMAL WHERE ASSET_ID='仪1-1151470269278326784';
+-----------------------------------+-------------------------+-------------+------------------------+--------------------------+----------------+
|                ID                 |        ASSET_ID         | ASSET_NAME  | ASSET_FIRST_DEGREE_ID  | ASSET_FIRST_DEGREE_NAME  | ASSET_SECOND_D |
+-----------------------------------+-------------------------+-------------+------------------------+--------------------------+----------------+
| 002e028151e24b07a21e0a0e9ce7f74c  | 仪1-  | 测量仪器        |                 | 仪表                       |      |
+-----------------------------------+-------------------------+-------------+------------------------+--------------------------+----------------+
 row selected (4.306 seconds)

参考:

https://my.oschina.net/puwenchao/blog/1935302

基于Phoenix对HBase建索引的更多相关文章

  1. hbase建索引的两种方式

    转载自http://blog.csdn.net/ryantotti/article/details/13295325 在二级索引的实现技术上一般有几个方案: 1.      表索引 使用单独的hbas ...

  2. Spark教程——(6)Spark-shell基于Phoenix访问HBase数据

    package statistics import common.util.timeUtil import org.apache.spark.{SparkConf, SparkContext} imp ...

  3. phoenix中添加二级索引

    Phoenix创建Hbase二级索引 官方文档 1. 配置Hbase支持Phoenix创建二级索引   1.  添加如下配置到Hbase的Hregionserver节点的hbase-site.xml  ...

  4. phoenix连接hbase数据库,创建二级索引报错:Error: org.apache.phoenix.exception.PhoenixIOException: Failed after attempts=36, exceptions: Tue Mar 06 10:32:02 CST 2018, null, java.net.SocketTimeoutException: callTimeou

    v\:* {behavior:url(#default#VML);} o\:* {behavior:url(#default#VML);} w\:* {behavior:url(#default#VM ...

  5. HBase之八--(2):HBase二级索引之Phoenix

    1. 介绍 Phoenix 是 Salesforce.com 开源的一个 Java 中间件,可以让开发者在Apache HBase 上执行 SQL 查询.Phoenix完全使用Java编写,代码位于 ...

  6. 「从零单排HBase 12」HBase二级索引Phoenix使用与最佳实践

    Phoenix是构建在HBase上的一个SQL层,能让我们用标准的JDBC APIs对HBase数据进行增删改查,构建二级索引.当然,开源产品嘛,自然需要注意“避坑”啦,阿丸会把使用方式和最佳实践都告 ...

  7. 通过phoenix在hbase上创建二级索引,Secondary Indexing

    环境描述: 操作系统版本:CentOS release 6.5 (Final) 内核版本:2.6.32-431.el6.x86_64 phoenix版本:phoenix-4.10.0 hbase版本: ...

  8. Hadoop生态圈-phoenix(HBase)的索引配置

    Hadoop生态圈-phoenix(HBase)的索引配置 作者:尹正杰 版权声明:原创作品,谢绝转载!否则将追究法律责任. 创建索引是为了优化查询,我们可以在phoenix上配置索引方式. 一.修改 ...

  9. Phoenix系列:二级索引(1)

    Phoenix使用HBase作为后端存储,对于HBase来说,我们通常使用字典序的RowKey来快速访问数据,除此之外,也可以使用自定义的Filter来搜索数据,但是它是基于全表扫描的.而Phoeni ...

随机推荐

  1. java 编译java文件 以及生成可执行jar

    1.新建java project: 2.src下新建包以及class文件: 3.打包: 5.选取目标mainclass 很关键决定jar是否可执行: 7.build jar : 8:artifact ...

  2. ES6-使用模板字符串完成字符串拼接

        var obj = {name:'tom',age:11};     //es5的字符串拼接比较麻烦     var str = '姓名是:'+obj.name+' '+'年龄是:'+obj. ...

  3. Python语言——map/reduce的用法

    Python内建了map()和reduce()函数. 如果你读过Google的那篇大名鼎鼎的论文“MapReduce: Simplified Data Processing on Large Clus ...

  4. 【PAT甲级】1064 Complete Binary Search Tree (30 分)

    题意:输入一个正整数N(<=1000),接着输入N个非负整数(<=2000),输出完全二叉树的层次遍历. AAAAAccepted code: #define HAVE_STRUCT_TI ...

  5. 最全BT磁力搜索引擎索引(整理分享,每日更新)

    btaa.xyz:http://www.veee.xyz/(可以访问,知名的BT磁力搜索,资源多,建议手机访问) 以下无法访问 idope.se:https://idope.se/(无法访问,资源丰富 ...

  6. 如何在PHP中防止SQL注入

    使用PDO对象(对于任何数据库驱动都好用) addslashes用于单字节字符串的处理, 多字节字符用mysql_real_escape_string吧. 另外对于php手册中get_magic_qu ...

  7. 《JavaScript高级程序设计》读书笔记(一)JavaScript简介

    起于客户端数据验证特性----闭包----匿名函数----元编程等----等想要全面理解和掌握JavaScript----本质----历史----局限性 ECMAScript 脚本语言标准 JavaS ...

  8. Java日期时间API系列9-----Jdk8中java.time包中的新的日期时间API类的Period和Duration的区别

    1.Period final修饰,线程安全,ISO-8601日历系统中基于日期的时间量,例如2年3个月4天. 主要属性:年数,月数,天数. /** * The number of years. */ ...

  9. Python经典排序算法

    https://www.cnblogs.com/onepixel/p/7674659.html这个文章很nice https://www.bilibili.com/video/av685670?fro ...

  10. python - 关于json和pickle两个序列化模块的区别

    传送门 https://stackoverflow.com/a/20980488/5955399 区别 json:用于字符串(unicode text)和python基本数据类型间进行转换.优点:跨语 ...