Phoenix詳細的信息可以查看： 官網 | apache/phoenix | Download | Performance | 支持的SQL語法
CDH6中集成的HBase版本爲2.1.0+cdh6.2.0 ,我們從官方的鏡像資源下載列表中看到最新的CDH支持到cdh5.14.2的版本，這種版本的直接有一個parcels包，可以通過Cloudera Manager的parcel頁面添加資源，來安裝組件。
通過查看源碼git branch -a可以看到 remotes/origin/5.x-cdh6 分支，但在源碼編譯安裝時這個分支還不完整，編譯出來時會缺少一些包，因此我們直接嘗試安裝v5.0.0-HBase-2.0版本的Phoenix，安裝方式有兩種，第一種直接下載官方編譯好的 apache-phoenix-5.0.0-HBase-2.0包apache-phoenix-5.0.0-HBase-2.0-bin.tar.gz；第二種是通過編譯源碼的方式來安裝。這裏我們直接使用源碼編譯來進行安裝。
目錄
- 1 Installation
- 1.1 源碼編譯
- 1.2 安裝
- 1.3 Cloudera Manager 的配置
- 2 Using
- 2.1 NoClassDefFoundError解決
- 2.2 原有HBase表的映射同步
- 2.3 查看錶信息和插入數據
- 2.4 數據類型
- 2.5 修改表信息
- 2.6 刪除庫
- 2.7 執行SQL腳本
- 2.8 DDL語法
- 2.9 DML和DQL語法
- 2.9.1 Join
- 2.10 JDBC
- 2.10.1 新建一個Maven工程
- 2.10.2 pom.xml
- 2.10.3 JDBC Client代碼
- 3 工具
- 3.1 DBeaver
- 3.2 SQuirrel SQL
- 4 資料
1 Installation
1.1 源碼編譯
# clone源碼
git clone https://github.com/apache/phoenix.git
# 進入源碼根目錄
cd phoenix/
# 查看tag ,或者查看分支 git branch -a
git tag
# 選擇一個版本，進入其分支，或者切換到某個分支：git checkout -b remotes/origin/5.x-cdh6
git checkout tags/v5.0.0-HBase-2.0
# 編譯
mvn -T2C install -DskipTests
如果編譯時報如下錯誤：
[INFO] ------------------------------------------------------------------------
[INFO] BUILD FAILURE
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 01:14 min (Wall Clock)
[INFO] Finished at: 2019-07-15T13:57:32+08:00
[INFO] ------------------------------------------------------------------------
[ERROR] Failed to execute goal net.alchim31.maven:scala-maven-plugin:3.2.0:compile (scala-compile-first) on project phoenix-spark: wrap: org.apache.commons.exec.ExecuteException: Process exited with an error: 1 (Exit value: 1) -> [Help 1]
[ERROR]
[ERROR] To see the full stack trace of the errors, re-run Maven with the -e switch.
[ERROR] Re-run Maven using the -X switch to enable full debug logging.
[ERROR]
[ERROR] For more information about the errors and possible solutions, please read the following articles:
[ERROR] [Help 1] http://cwiki.apache.org/confluence/display/MAVEN/MojoExecutionException
[ERROR]
[ERROR] After correcting the problems, you can resume the build with the command
[ERROR] mvn <goals> -rf :phoenix-spark
可以看到是phoenix-spark模塊的net.alchim31.maven:scala-maven-plugin:3.2.0插件編譯時發生了錯誤，我們可以將其原有的scala-maven-plugin註釋掉，修改爲如下，然後重新編譯：
<plugin>
<groupId>net.alchim31.maven</groupId>
<artifactId>scala-maven-plugin</artifactId>
<version>3.2.2</version>
<executions>
<execution>
<goals>
<goal>compile</goal>
<goal>testCompile</goal>
</goals>
</execution>
</executions>
</plugin>
編譯成功後如下：
[INFO] ------------------------------------------------------------------------
[INFO] Reactor Summary for Apache Phoenix 5.0.0-HBase-2.0:
[INFO]
[INFO] Apache Phoenix ..................................... SUCCESS [ 9.616 s]
[INFO] Phoenix Core ....................................... SUCCESS [03:34 min]
[INFO] Phoenix - Flume .................................... SUCCESS [01:56 min]
[INFO] Phoenix - Kafka .................................... SUCCESS [01:33 min]
[INFO] Phoenix - Pig ...................................... SUCCESS [04:59 min]
[INFO] Phoenix Query Server Client ........................ SUCCESS [ 48.315 s]
[INFO] Phoenix Query Server ............................... SUCCESS [07:59 min]
[INFO] Phoenix - Pherf .................................... SUCCESS [ 8.362 s]
[INFO] Phoenix - Spark .................................... SUCCESS [09:20 min]
[INFO] Phoenix - Hive ..................................... SUCCESS [57:28 min]
[INFO] Phoenix Client ..................................... SUCCESS [01:20 min]
[INFO] Phoenix Server ..................................... SUCCESS [ 37.906 s]
[INFO] Phoenix Assembly ................................... SUCCESS [ 19.902 s]
[INFO] Phoenix - Tracing Web Application .................. SUCCESS [ 9.255 s]
[INFO] Phoenix Load Balancer .............................. SUCCESS [ 19.176 s]
[INFO] ------------------------------------------------------------------------
[INFO] BUILD SUCCESS
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 01:01 h (Wall Clock)
[INFO] Finished at: 2019-07-07T11:09:10+08:00
[INFO] ------------------------------------------------------------------------
1.2 安裝
將編譯成功的phoenix-assembly/target/phoenix-5.0.0-HBase-2.0.tar.gz包拷貝到Phoenix的安裝節點，然後解壓
cp phoenix-assembly/target/phoenix-5.0.0-HBase-2.0.tar.gz /opt/
cd /opt
tar -zxf phoenix-5.0.0-HBase-2.0.tar.gz
cd phoenix-5.0.0-HBase-2.0
此時爲了方便可以配置下環境變量
vim ~/.bash_profile
# 添加如下配置，並保存退出
export PHOENIX_HOME=/opt/phoenix-5.0.0-HBase-2.0
export PATH=$PATH:$PHOENIX_HOME/bin
接下來將Phoenix根目錄下的phoenix-5.0.0-HBase-2.0-*.jar 包分發到每個RegionServer的lib下
scp phoenix-5.0.0-HBase-2.0-*.jar root@cdh3:/opt/cloudera/parcels/CDH-6.2.0-1.cdh6.2.0.p0.967373/lib/hbase/lib/
scp phoenix-5.0.0-HBase-2.0-*.jar root@cdh2:/opt/cloudera/parcels/CDH-6.2.0-1.cdh6.2.0.p0.967373/lib/hbase/lib/
scp phoenix-5.0.0-HBase-2.0-*.jar root@cdh1:/opt/cloudera/parcels/CDH-6.2.0-1.cdh6.2.0.p0.967373/lib/hbase/lib/
這裏最好也將phoenix-5.0.0-HBase-2.0-client.jar拷貝到Phoenix安裝的根目錄下
cp phoenix-client/target/phoenix-5.0.0-HBase-2.0-client.jar $PHOENIX_HOME/
1.3 Cloudera Manager 的配置
登陸Cloudera Manager頁面，然後修改HBase配置，進入配置頁面後搜索hbase-site.xml 的 HBase 服務高級配置代碼段，然後添加如下配置，然後重啓 HBase 使配置生效。
<property>
<name>hbase.table.sanity.checks</name>
<value>false</value>
</property>
2 Using
進入Phoenix的命令行執行
# 後面跟的參數爲HBase配置的ZK的地址
sqlline.py cdh2:2181
2.1 NoClassDefFoundError解決
如果進入Phoenix終端時報如下錯誤：
Error: org.apache.phoenix.exception.PhoenixIOException: org.apache.hadoop.hbase.DoNotRetryIOException: java.lang.NoClassDefFoundError: org/apache/htrace/Trace
at org.apache.hadoop.hbase.ipc.RpcServer.call(RpcServer.java:469)
at org.apache.hadoop.hbase.ipc.CallRunner.run(CallRunner.java:130)
at org.apache.hadoop.hbase.ipc.RpcExecutor$Handler.run(RpcExecutor.java:324)
at org.apache.hadoop.hbase.ipc.RpcExecutor$Handler.run(RpcExecutor.java:304)
Caused by: java.lang.NoClassDefFoundError: org/apache/htrace/Trace
at org.apache.phoenix.coprocessor.BaseScannerRegionObserver$RegionScannerHolder.overrideDelegate(BaseScannerRegionObserver.java:222)
at org.apache.phoenix.coprocessor.BaseScannerRegionObserver$RegionScannerHolder.nextRaw(BaseScannerRegionObserver.java:273)
at org.apache.hadoop.hbase.regionserver.RSRpcServices.scan(RSRpcServices.java:3134)
at org.apache.hadoop.hbase.regionserver.RSRpcServices.scan(RSRpcServices.java:3383)
at org.apache.hadoop.hbase.shaded.protobuf.generated.ClientProtos$ClientService$2.callBlockingMethod(ClientProtos.java:42002)
at org.apache.hadoop.hbase.ipc.RpcServer.call(RpcServer.java:413)
... 3 more
Caused by: java.lang.ClassNotFoundException: org.apache.htrace.Trace
at java.net.URLClassLoader.findClass(URLClassLoader.java:382)
at java.lang.ClassLoader.loadClass(ClassLoader.java:424)
at sun.misc.Launcher$AppClassLoader.loadClass(Launcher.java:349)
at java.lang.ClassLoader.loadClass(ClassLoader.java:357)
... 9 more (state=08000,code=101)
可以看到提示缺少org.apache.htrace.Trace類，因此我們需要將htrace-core-3.2.0-incubating.jar拷貝到HBase的lib下：
# CDH環境下已經有這個包了，這裏直接將其拷貝到HBase的lib下（每個RegionServer）
# 如果沒有可以直接下載這包到HBase的lib下： wget https://repo1.maven.org/maven2/org/apache/htrace/htrace-core/3.2.0-incubating/htrace-core-3.2.0-incubating.jar
cp /opt/cloudera/cm/lib/cdh5/htrace-core-3.2.0-incubating.jar /opt/cloudera/parcels/CDH-6.2.0-1.cdh6.2.0.p0.967373/lib/hbase/lib/
注意： 如果依舊提示缺少這個包，則需要重啓HBase服務
再次進入
[root@cdh3 phoenix-5.0.0-HBase-2.0]# bin/sqlline.py cdh2
Setting property: [incremental, false]
Setting property: [isolation, TRANSACTION_READ_COMMITTED]
issuing: !connect jdbc:phoenix:cdh2 none none org.apache.phoenix.jdbc.PhoenixDriver
Connecting to jdbc:phoenix:cdh2
SLF4J: Class path contains multiple SLF4J bindings.
SLF4J: Found binding in [jar:file:/opt/phoenix-5.0.0-HBase-2.0/phoenix-5.0.0-HBase-2.0-client.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: Found binding in [jar:file:/opt/cloudera/parcels/CDH-6.2.0-1.cdh6.2.0.p0.967373/jars/slf4j-log4j12-1.7.25.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: See http://www.slf4j.org/codes.html#multiple_bindings for an explanation.
Connected to: Phoenix (version 5.0)
Driver: PhoenixEmbeddedDriver (version 5.0)
Autocommit status: true
Transaction isolation: TRANSACTION_READ_COMMITTED
Building list of tables and columns for tab-completion (set fastconnect to true to skip)...
133/133 (100%) Done
Done
sqlline version 1.2.0
0: jdbc:phoenix:cdh2> !tables
+------------+--------------+-------------+---------------+----------+------------+----------------------------+-----------------+--------------+-----------------+---------------+---------------+-----------------+------------+-------------+----------------+-------------+
| TABLE_CAT | TABLE_SCHEM | TABLE_NAME | TABLE_TYPE | REMARKS | TYPE_NAME | SELF_REFERENCING_COL_NAME | REF_GENERATION | INDEX_STATE | IMMUTABLE_ROWS | SALT_BUCKETS | MULTI_TENANT | VIEW_STATEMENT | VIEW_TYPE | INDEX_TYPE | TRANSACTIONAL | IS_NAMESPAC |
+------------+--------------+-------------+---------------+----------+------------+----------------------------+-----------------+--------------+-----------------+---------------+---------------+-----------------+------------+-------------+----------------+-------------+
| | SYSTEM | CATALOG | SYSTEM TABLE | | | | | | false | null | false | | | | false | false |
| | SYSTEM | FUNCTION | SYSTEM TABLE | | | | | | false | null | false | | | | false | false |
| | SYSTEM | LOG | SYSTEM TABLE | | | | | | true | 32 | false | | | | false | false |
| | SYSTEM | SEQUENCE | SYSTEM TABLE | | | | | | false | null | false | | | | false | false |
| | SYSTEM | STATS | SYSTEM TABLE | | | | | | false | null | false | | | | false | false |
+------------+--------------+-------------+---------------+----------+------------+----------------------------+-----------------+--------------+-----------------+---------------+---------------+-----------------+------------+-------------+----------------+-------------+
2.2 原有HBase表的映射同步
Phoenix對錶和列名都是區分大小寫的，如果不加雙引號默認爲大寫。通過上面可以看到HBase原有的表並沒有列出來，這是因爲 Phoenix 無法自動識別 HBase 中原有的表，所以需要將 HBase 中已有的做映射，才能夠被 Phoenix 識別並操作。
-- 其實也就是使用CREATE語句重新創建表來映射
-- 注意1：建表時一定注意大小寫，如果原表小寫，記得用雙引號，主鍵最好大寫
-- 注意2：建表時不推薦使用太複雜的數據類型，爲了安全的映射放置出錯，先使用varchar，
-- 否則會出現精度或者格式上的誤差，例如時間日期類，整數型長度，小數型精度等
-- 注意3：記得加上 COLUMN_ENCODED_BYTES 否則可能類名解析無法對應到HBase的類，
-- 詳細說明查看 http://phoenix.apache.org/columnencoding.html
-- 注意4：如果是創建的表來映射，在Phoenix刪除表，HBase的表和數據也會刪除，注意數據不要誤刪了
--
-- 推薦：我這裏映射的是表，如果只用來做查詢，最好創建一個視圖來來映射：只需要將 TABLE 改爲 VIEW
0: jdbc:phoenix:cdh2:2181> CREATE TABLE "m_info" (
. . . . . . . . . . . . .> "ROW" varchar primary key,
. . . . . . . . . . . . .> "cf1"."gc" double ,
. . . . . . . . . . . . .> "cf1"."id" bigint ,
. . . . . . . . . . . . .> "cf1"."jd" double ,
. . . . . . . . . . . . .> "cf1"."name" varchar ,
. . . . . . . . . . . . .> "cf1"."wd" double
. . . . . . . . . . . . .> )column_encoded_bytes=0;
1 row affected (7.282 seconds)
-- 查詢數據
0: jdbc:phoenix:cdh2> select * from "m_info" ;
+------+-----------------------+-----------------------+-----------------------+-------+-------------------------+
| ROW | gc | id | jd | name | wd |
+------+-----------------------+-----------------------+-----------------------+-------+-------------------------+
| 2 | 1.1 | 2 | 3.33 | 名字1 | 5.1234 |
+------+-----------------------+-----------------------+-----------------------+-------+-------------------------+
2.3 查看錶信息和插入數據
-- 查看錶結構
0: jdbc:phoenix:cdh2:2181> !describe "m_info";
+------------+--------------+-------------+--------------+------------+------------+--------------+----------------+-----------------+-----------------+-----------+----------+-------------+----------------+-------------------+--------------------+-------------------+---+
| TABLE_CAT | TABLE_SCHEM | TABLE_NAME | COLUMN_NAME | DATA_TYPE | TYPE_NAME | COLUMN_SIZE | BUFFER_LENGTH | DECIMAL_DIGITS | NUM_PREC_RADIX | NULLABLE | REMARKS | COLUMN_DEF | SQL_DATA_TYPE | SQL_DATETIME_SUB | CHAR_OCTET_LENGTH | ORDINAL_POSITION | I |
+------------+--------------+-------------+--------------+------------+------------+--------------+----------------+-----------------+-----------------+-----------+----------+-------------+----------------+-------------------+--------------------+-------------------+---+
| | | m_info | ROW | 12 | VARCHAR | null | null | null | null | 0 | | | null | null | null | 1 | f |
| | | m_info | gc | 8 | DOUBLE | null | null | null | null | 1 | | | null | null | null | 2 | t |
| | | m_info | id | -5 | BIGINT | null | null | null | null | 1 | | | null | null | null | 3 | t |
| | | m_info | jd | 8 | DOUBLE | null | null | null | null | 1 | | | null | null | null | 4 | t |
| | | m_info | name | 12 | VARCHAR | null | null | null | null | 1 | | | null | null | null | 5 | t |
| | | m_info | wd | 8 | DOUBLE | null | null | null | null | 1 | | | null | null | null | 6 | t |
+------------+--------------+-------------+--------------+------------+------------+--------------+----------------+-----------------+-----------------+-----------+----------+-------------+----------------+-------------------+--------------------+-------------------+---+
-- 插入數據。這裏注意的是value中的值如果是字符串，必須用單引號
0: jdbc:phoenix:cdh2:2181> UPSERT INTO "m_info"("ROW", "gc", "id", "jd", "name", "wd") VALUES('2', 1.1, 2, 3.33, '名字1', 5.1234);
1 row affected (1.074 seconds)
2.4 數據類型
| 字段類型 | 映射的類 | 值的範圍 | 佔用字節 |
|---|---|---|---|
| INTEGER | java.lang.Integer |
-2147483648 到 2147483647 | 4字節 |
| UNSIGNED_INT | java.lang.Integer |
0 到 2147483647 | 4字節 |
| BIGINT | java.lang.Long |
-9223372036854775808 到 9223372036854775807 | 8字節 |
| UNSIGNED_LONG | java.lang.Long |
0 到 9223372036854775807 | 8字節 |
| TINYINT | java.lang.Byte |
-128 到 127 | 1字節 |
| UNSIGNED_TINYINT | java.lang.Byte |
0 到 127 | 1字節 |
| SMALLINT | java.lang.Short |
-32768 到 32767 | 2字節 |
| UNSIGNED_SMALLINT | java.lang.Short |
0 到 32767 | 2字節 |
| FLOAT | java.lang.Float |
-3.402823466 E + 38 到 3.402823466 E + 38 | 4字節 |
| UNSIGNED_FLOAT | java.lang.Float |
0到3.402823466 E + 38 | 4字節 |
| DOUBLE | java.lang.Double |
-1.7976931348623158 E + 308 到 1.7976931348623158 E + 308 | 8字節 |
| UNSIGNED_DOUBLE | java.lang.Double |
0 到 1.7976931348623158 E + 308 | 8字節 |
| DECIMAL 或 DECIMAL(precisionInt , scaleInt) | java.math.BigDecimal |
最大精度爲38位。可變長度 | |
| BOOLEAN | java.lang.Boolean |
0表示false，1表示true | 1字節 |
| TIME | java.sql.Time |
自紀元以來的毫秒數（基於時間GMT） | 8字節 |
| UNSIGNED_TIME | java.sql.Time |
格式爲 yyyy-MM-dd hh:mm:ss | 8字節 |
| DATE | java.sql.Date |
自紀元以來的毫秒數（基於時間GMT） | 8字節 |
| UNSIGNED_DATE | java.sql.Date |
格式爲 yyyy-MM-dd hh:mm:ss | 8字節 |
| TIMESTAMP | java.sql.Timestamp |
自紀元以來的毫秒數（基於時間GMT） | 12字節（紀元時間長8字節加納秒的4字節整數） |
| UNSIGNED_TIMESTAMP | java.sql.Timestamp |
格式爲 yyyy-MM-dd hh:mm:ss[.nnnnnnnnn] | 12字節 |
| VARCHAR 或 VARCHAR(precisionInt) | java.lang.String |
可選的最大字節長度 | |
| CHAR(precisionInt) | java.lang.String |
固定長度字符串 | |
| BINARY(precisionInt) | byte[] |
原始固定長度字節數組 | |
| VARBINARY | byte[] |
原始可變長度字節數組 | |
| ARRAY 或 ARRAY[dimensionInt] | java.sql.Array |
2.5 修改表信息
首先Phoenix對字段的映射的元數據信息如下，下面修改字段類型時可以由此進行修改
| DATA_TYPE | TYPE_NAME |
|---|---|
| INTEGER | 4 |
| UNSIGNED_INT | 4 |
| BIGINT | -5 |
| UNSIGNED_LONG | -5 |
| TINYINT | -6 |
| UNSIGNED_TINYINT | -6 |
| SMALLINT | 5 |
| UNSIGNED_SMALLINT | 5 |
| FLOAT | 6 |
| UNSIGNED_FLOAT | 6 |
| DOUBLE | 8 |
| UNSIGNED_DOUBLE | 8 |
| DECIMAL | 3 |
| BOOLEAN | 16 |
| TIME | 92 |
| UNSIGNED_TIME | 92 |
| DATE | 91 |
| UNSIGNED_DATE | 91 |
| TIMESTAMP | 93 |
| UNSIGNED_TIMESTAMP | 93 |
| VARCHAR | 12 |
| CHAR | 1 |
| BINARY | -2 |
| VARBINARY | -3 |
0: jdbc:phoenix:cdh2> CREATE TABLE "test" (
. . . . . . . . . . > "ROW" varchar primary key,
. . . . . . . . . . > "info1"."age" UNSIGNED_TINYINT ,
. . . . . . . . . . > "info1"."name" varchar ,
. . . . . . . . . . > "info1"."property" double
. . . . . . . . . . > )column_encoded_bytes=0;
3 rows affected (8.962 seconds)
-- 查看錶中的字段信息
0: jdbc:phoenix:cdh2> select TENANT_ID,TABLE_SCHEM,TABLE_NAME,COLUMN_NAME,COLUMN_FAMILY,DATA_TYPE,COLUMN_SIZE,DECIMAL_DIGITS from SYSTEM.CATALOG where TABLE_NAME='test';
+------------+--------------+-------------+--------------+----------------+------------+--------------+-----------------+
| TENANT_ID | TABLE_SCHEM | TABLE_NAME | COLUMN_NAME | COLUMN_FAMILY | DATA_TYPE | COLUMN_SIZE | DECIMAL_DIGITS |
+------------+--------------+-------------+--------------+----------------+------------+--------------+-----------------+
| | | test | | | null | null | null |
| | | test | ROW | | 12 | null | null |
| | | test | age | info1 | -6 | null | null |
| | | test | name | info1 | 12 | null | null |
| | | test | property | info1 | 8 | null | null |
+------------+--------------+-------------+--------------+----------------+------------+--------------+-----------------+
5 rows selected (0.064 seconds)
-- 將name的varchar類型的長度調整爲 32，前4個字段爲主鍵
5 rows selected (0.064 seconds)
0: jdbc:phoenix:cdh2> upsert into SYSTEM.CATALOG (TENANT_ID,TABLE_SCHEM,TABLE_NAME,COLUMN_NAME,COLUMN_FAMILY,COLUMN_SIZE) values('','','test','name','info1',32);
1 row affected (0.493 seconds)
0: jdbc:phoenix:cdh2> upsert into SYSTEM.CATALOG (TENANT_ID,TABLE_SCHEM,TABLE_NAME,COLUMN_NAME,COLUMN_FAMILY,DATA_TYPE) values('','','test','age','info1',-5);
1 row affected (0.019 seconds)
0: jdbc:phoenix:cdh2> select TENANT_ID,TABLE_SCHEM,TABLE_NAME,COLUMN_NAME,COLUMN_FAMILY,DATA_TYPE,COLUMN_SIZE,DECIMAL_DIGITS from SYSTEM.CATALOG where TABLE_NAME='test';
+------------+--------------+-------------+--------------+----------------+------------+--------------+-----------------+
| TENANT_ID | TABLE_SCHEM | TABLE_NAME | COLUMN_NAME | COLUMN_FAMILY | DATA_TYPE | COLUMN_SIZE | DECIMAL_DIGITS |
+------------+--------------+-------------+--------------+----------------+------------+--------------+-----------------+
| | | test | | | null | null | null |
| | | test | ROW | | 12 | null | null |
| | | test | age | info1 | -5 | null | null |
| | | test | name | info1 | 12 | 32 | null |
| | | test | property | info1 | 8 | null | null |
+------------+--------------+-------------+--------------+----------------+------------+--------------+-----------------+
5 rows selected (0.039 seconds)
2.6 刪除庫
-- 如果HBase的數據和Phoenix的數據是同步的，則可以直接刪除SCHEM
0: jdbc:phoenix:cdh3> DROP SCHEMA IF EXISTS my_schema;
-- 如果不能刪除，可以先刪除HBase的表信息
hbase(main):003:0> disable "test.Person"
Took 0.5061 seconds
hbase(main):004:0> drop "test.Person"
Took 0.2915 seconds
-- 再清除Phoenix的數據
-- 查看CATALOG信息
0: jdbc:phoenix:cdh3> select TENANT_ID,TABLE_SCHEM,TABLE_NAME,COLUMN_NAME,COLUMN_FAMILY,TABLE_SEQ_NUM,TABLE_TYPE,PK_NAME from SYSTEM.CATALOG;
-- 刪除裏面的表信息
| TENANT_ID | TABLE_SCHEM | TABLE_NAME | COLUMN_NAME | COLUMN_FAMILY | TABLE_SEQ_NUM | TABLE_TYPE | PK_NAME |
+------------+--------------+-------------+-------------------------------------+----------------+----------------+-------------+----------+
| | …… | …… | …… | | null | | |
| | test | Person | | | 0 | u | |
| | test | Person | | 0 | null | | |
| | test | Person | age | 0 | null | | |
| | test | Person | idcard_num | | null | | |
| | test | Person | name | 0 | null | | |
+------------+--------------+-------------+------------------------+----------------+----------------+-------------+----------+
-- 刪除SCHEM信息
0: jdbc:phoenix:cdh3> DELETE FROM SYSTEM.CATALOG WHERE TABLE_SCHEM='test';
5 rows affected (0.352 seconds)
-- 再次使用 !tables 查看錶，發現庫以及表均已經成功刪除
2.7 執行SQL腳本
我們使用Phoenix自帶的樣例數據演示，數據在 $PHOENIX_HOME/examples，具體執行如下：
-- 1 查看數據和腳本，其中 $PHOENIX_HOME 是Phoenix安裝的目錄。
-- 可以看到有STOCK_SYMBOL數據文件和sql腳本、WEB_STAT數據文件和腳本文件
[root@cdh3 ~]# ls $PHOENIX_HOME/examples
pig STOCK_SYMBOL.csv STOCK_SYMBOL.sql WEB_STAT.csv WEB_STAT_QUERIES.sql WEB_STAT.sql
-- 1.1 查看 STOCK_SYMBOL.sql。可以看到有三個SQL，分別爲建表SQL、更新插入一條數據SQL、查詢表數據SQL
[root@cdh3 ~]# cat $PHOENIX_HOME/examples/STOCK_SYMBOL.sql
CREATE TABLE IF NOT EXISTS STOCK_SYMBOL (SYMBOL VARCHAR NOT NULL PRIMARY KEY, COMPANY VARCHAR);
UPSERT INTO STOCK_SYMBOL VALUES ('CRM','SalesForce.com');
SELECT * FROM STOCK_SYMBOL;
-- 1.2 查看 WEB_STAT.sql。可以看到這裏是一個建表SQL
[root@cdh3 ~]# cat $PHOENIX_HOME/examples/WEB_STAT.sql
CREATE TABLE IF NOT EXISTS WEB_STAT (
HOST CHAR(2) NOT NULL,
DOMAIN VARCHAR NOT NULL,
FEATURE VARCHAR NOT NULL,
DATE DATE NOT NULL,
USAGE.CORE BIGINT,
USAGE.DB BIGINT,
STATS.ACTIVE_VISITOR INTEGER
CONSTRAINT PK PRIMARY KEY (HOST, DOMAIN, FEATURE, DATE)
);
-- 1.3 查看 WEB_STAT_QUERIES.sql。可以看到有三個查詢 SQL
[root@cdh3 ~]# cat $PHOENIX_HOME/examples/WEB_STAT_QUERIES.sql
SELECT DOMAIN, AVG(CORE) Average_CPU_Usage, AVG(DB) Average_DB_Usage
FROM WEB_STAT
GROUP BY DOMAIN
ORDER BY DOMAIN DESC;
-- Sum, Min and Max CPU usage by Salesforce grouped by day
SELECT TRUNC(DATE,'DAY') DAY, SUM(CORE) TOTAL_CPU_Usage, MIN(CORE) MIN_CPU_Usage, MAX(CORE) MAX_CPU_Usage
FROM WEB_STAT
WHERE DOMAIN LIKE 'Salesforce%'
GROUP BY TRUNC(DATE,'DAY');
-- list host and total active users when core CPU usage is 10X greater than DB usage
SELECT HOST, SUM(ACTIVE_VISITOR) TOTAL_ACTIVE_VISITORS
FROM WEB_STAT
WHERE DB > (CORE * 10)
GROUP BY HOST;
-- 2 執行自帶腳本
[root@cdh3 ~]# $PHOENIX_HOME/bin/sqlline.py localhost $PHOENIX_HOME/examples/STOCK_SYMBOL.sql
148/148 (100%) Done
Done
1/4 CREATE TABLE IF NOT EXISTS STOCK_SYMBOL (SYMBOL VARCHAR NOT NULL PRIMARY KEY, COMPANY VARCHAR);
No rows affected (2.349 seconds)
2/4 UPSERT INTO STOCK_SYMBOL VALUES ('CRM','SalesForce.com');
1 row affected (0.162 seconds)
3/4 SELECT * FROM STOCK_SYMBOL;
+---------+-----------------+
| SYMBOL | COMPANY |
+---------+-----------------+
| CRM | SalesForce.com |
+---------+-----------------+
1 row selected (0.038 seconds)
4/4
Closing: org.apache.phoenix.jdbc.PhoenixConnection
sqlline version 1.2.0
-- 也可以指定csv數據文件和執行sql腳本
[root@cdh3 ~]# psql.py localhost $PHOENIX_HOME/examples/WEB_STAT.sql $PHOENIX_HOME/examples/WEB_STAT.csv $PHOENIX_HOME/examples/WEB_STAT_QUERIES.sql
no rows upserted
Time: 2.709 sec(s)
csv columns from database.
CSV Upsert complete. 39 rows upserted
Time: 19.653 sec(s)
DOMAIN AVERAGE_CPU_USAGE AVERAGE_DB_USAGE
----------------------------- ------------------------------- --------------------------
Salesforce.com 260.727 257.636
Google.com 212.875 213.75
Apple.com 114.111 119.556
Time: 0.127 sec(s)
DAY TOTAL_CPU_USAGE MIN_CPU_USAGE MAX_CPU_USAGE
----------------------- -------------------- -------------------------- ----------------------------
2013-01-01 00:00:00.000 35 35 35
2013-01-02 00:00:00.000 150 25 125
2013-01-03 00:00:00.000 88 88 88
2013-01-04 00:00:00.000 26 3 23
2013-01-05 00:00:00.000 550 75 475
2013-01-06 00:00:00.000 12 12 12
2013-01-08 00:00:00.000 345 345 345
2013-01-09 00:00:00.000 390 35 355
2013-01-10 00:00:00.000 345 345 345
2013-01-11 00:00:00.000 335 335 335
2013-01-12 00:00:00.000 5 5 5
2013-01-13 00:00:00.000 355 355 355
2013-01-14 00:00:00.000 5 5 5
2013-01-15 00:00:00.000 720 65 655
2013-01-16 00:00:00.000 785 785 785
2013-01-17 00:00:00.000 1590 355 1235
Time: 0.045 sec(s)
HO TOTAL_ACTIVE_VISITORS
-- ----------------------------------------
EU 150
NA 1
Time: 0.022 sec(s)
-- 3 登陸 Phoenix。也可以使用Squirrel工具
[root@cdh3 ~]# $PHOENIX_HOME/bin/sqlline.py localhost
2.8 DDL語法
數據庫定義語言DDL（Data Definition Language）是我們在創建表的時候用到的一些SQL，比如說：CREATE、ALTER、DROP等。DDL主要是用在定義或改變表的結構，數據類型，表之間的鏈接和約束等初始化工作上。
-- 1 建表，建表的語句可以看 執行SQL腳本 中的如下兩個文件的建表語句
-- 基本上和一般的SQL建表差不多，支持主鍵，支持 NOT NULL約束，這裏又不同的是，建表時可以在字段前加上列族名以區分字段信息
$PHOENIX_HOME/examples/STOCK_SYMBOL.sql
$PHOENIX_HOME/examples/WEB_STAT.sql
-- 2 ALTER一個表，例如在STOCK_SYMBOL添加一列 name
0: jdbc:phoenix:localhost> ALTER TABLE STOCK_SYMBOL ADD "name" varchar(64);
No rows affected (0.749 seconds)
-- 再次查看錶數據，發現有多了一列name
0: jdbc:phoenix:localhost> SELECT * FROM STOCK_SYMBOL;
+---------+-----------------+-------+
| SYMBOL | COMPANY | name |
+---------+-----------------+-------+
| CRM | SalesForce.com | |
+---------+-----------------+-------+
3 rows selected (0.15 seconds)
-- 3 ALTER一個表，刪除一列，例如將上一步在STOCK_SYMBOL表添加的name一列刪除
0: jdbc:phoenix:localhost> ALTER TABLE STOCK_SYMBOL DROP COLUMN "name";
No rows affected (0.605 seconds)
-- 再次查看錶數據，發現name這一列已經刪除
0: jdbc:phoenix:localhost> SELECT * FROM STOCK_SYMBOL;
+---------+-----------------+
| SYMBOL | COMPANY |
+---------+-----------------+
| CRM | SalesForce.com |
+---------+-----------------+
3 rows selected (0.095 seconds)
-- 如果demo2 表是一個視圖，也可以將視圖中的某個列字段刪去，語法如下
0: jdbc:phoenix:localhost> ALTER VIEW "demo2" DROP COLUMN "comm";
-- 4 刪除庫
-- 4.1 如果HBase的數據和Phoenix的數據是同步的，則可以直接刪除SCHEM
0: jdbc:phoenix:cdh3> DROP SCHEMA IF EXISTS my_schema;
-- 4.2 如果不能刪除，可以先刪除HBase的表信息
hbase(main):003:0> disable "test.Person"
Took 0.5061 seconds
hbase(main):004:0> drop "test.Person"
Took 0.2915 seconds
-- 再清除Phoenix的數據
-- 4.3 刪除SCHEM信息
0: jdbc:phoenix:cdh3> DELETE FROM SYSTEM.CATALOG WHERE TABLE_SCHEM='test';
5 rows affected (0.352 seconds)
-- 再次使用 !tables 查看錶，發現庫以及表均已經成功刪除
2.8.1 加鹽表
在密碼學中，加鹽是指在散列之前將散列內容（例如：密碼）的任意固定位置插入特定的字符串。這個在散列中加入字符串的方式稱爲“加鹽”。其作用是讓加鹽後的散列結果和沒有加鹽的結果不相同，在不同的應用情景中，這個處理可以增加額外的安全性。而Phoenix中加鹽是指對pk對應的byte數組插入特定的byte數據。
加鹽能解決HBASE讀寫熱點問題，例如:單調遞增rowkey數據的持續寫入，使得負載集中在某一個RegionServer上引起的熱點問題。在創建表的時候指定屬性值：SALT_BUCKETS，其值表示所分buckets(region)數量， 範圍是1~256。
-- 加鹽表的創建格式如下
CREATE TABLE mytable (my_key VARCHAR PRIMARY KEY, col VARCHAR) SALT_BUCKETS = 8;
2.8.2 自增ID
現在有一個需求是這樣的，有一批書籍，需要對其進行編號存儲，要求編號要唯一，建表語句如下：
0: jdbc:phoenix:localhost> CREATE TABLE books(
. . . . . . . . . . . . .> id INTEGER NOT NULL PRIMARY KEY,
. . . . . . . . . . . . .> name VARCHAR,
. . . . . . . . . . . . .> author VARCHAR
. . . . . . . . . . . . .> )SALT_BUCKETS = 8;
No rows affected (3.250 seconds)
我們接下來需要使用自增ID，來生成上表的id，但是使用這樣的自增ID，容易造成集羣的熱點問題，因此這裏在建表時通過加鹽的方式解決這個潛在的問題。下面是Phoenix中自增ID的語法，如下：
-- 自增ID的語法如下：
-- START 用於指定第一個值。如果不指定默認爲1.
-- INCREMENT指定每次調用NEXT VALUE FOR後自增大小。 如果不指定默認爲1。
-- MINVALUE和MAXVALUE一般與CYCLE連用, 讓自增數據形成一個環，從最小值到最大值，再從最大值到最小值。
-- CACHE默認爲100, 表示server端生成100個自增序列緩存在客戶端，可以減少rpc次數。
-- 此值也可以通過phoenix.sequence.cacheSize來配置。
CREATE SEQUENCE [IF NOT EXISTS] SCHEMA.SEQUENCE_NAME
[START WITH number]
[INCREMENT BY number]
[MINVALUE number]
[MAXVALUE number]
[CYCLE]
[CACHE number]
現在我們使用這個語法創建自增序列，我們定初始值爲10000，自增間隔爲1，緩存大小爲1000，實現如下：
-- 1 創建自增序列
0: jdbc:phoenix:localhost> CREATE SEQUENCE book_sequence START WITH 10000 INCREMENT BY 1 CACHE 1000;
No rows affected (0.174 seconds)
-- 2 通過自增序列，寫入數據信息。這裏插入兩條數據，通過NEXT VALUE FOR 調用自增序列
0: jdbc:phoenix:localhost> UPSERT INTO books(id, name, author) VALUES( NEXT VALUE FOR
. . . . . . . . . . . . .> book_sequence,'Effective JavaScript', 'David Herman');
1 row affected (0.085 seconds)
-- 插入第二條數據
0: jdbc:phoenix:localhost> UPSERT INTO books(id, name, author) VALUES( NEXT VALUE FOR
. . . . . . . . . . . . .> book_sequence,'Speaking Javascript','Dr.Axel Rauschmayer');
1 row affected (0.014 seconds)
-- 3 查看書籍信息表。books沒有加引號時默認爲大寫。發現數據已成功插入，且其ID爲從10000以1位步長自增
0: jdbc:phoenix:localhost> SELECT * FROM BOOKS;
+--------+-----------------------+----------------------+
| ID | NAME | AUTHOR |
+--------+-----------------------+----------------------+
| 10001 | Speaking Javascript | Dr.Axel Rauschmayer |
| 10000 | Effective JavaScript | David Herman |
+--------+-----------------------+----------------------+
2 rows selected (0.245 seconds)
2.9 DML和DQL語法
數據操縱語言DML（Data Manipulation Language）主要對錶中的數據進行插入、刪除、修改的操作，在Phoenix對應的操作就是UPSERT和DELETE。數據查詢語言DQL（Data Query Language）主要用於數據的查詢，在Phoenix對應的操作就是SELETE。
-- 1 SELECT 。查詢 STOCK_SYMBOL 表數據，取其3條
0: jdbc:phoenix:localhost> SELECT * FROM WEB_STAT LIMIT 3;
+-------+-------------+------------+--------------------------+-------+------+-----------------+
| HOST | DOMAIN | FEATURE | DATE | CORE | DB | ACTIVE_VISITOR |
+-------+-------------+------------+--------------------------+-------+------+-----------------+
| EU | Apple.com | Mac | 2013-01-01 01:01:01.000 | 35 | 22 | 34 |
| EU | Apple.com | Store | 2013-01-03 01:01:01.000 | 345 | 722 | 170 |
| EU | Google.com | Analytics | 2013-01-13 08:06:01.000 | 25 | 2 | 6 |
+-------+-------------+------------+--------------------------+-------+------+-----------------+
3 rows selected (0.132 seconds)
-- 2 UPSERT。這個是需要注意的，因爲HBase本身的架構特點，沒有update語法，Phoenix的插入和修改統一爲 UPSERT 關鍵字。
-- 2.1 插入一條數據，不指定字段，直接填寫插入的值，建表時有多少字段就要一一對應，不能漏
0: jdbc:phoenix:localhost> UPSERT INTO STOCK_SYMBOL VALUES('AAPL', 'APPLE Inc.');
1 row affected (0.355 seconds)
-- 2.2 插入一條數據，可以只指定我們需要插入數據的字段，其它沒有約束時使用默認值
0: jdbc:phoenix:localhost> UPSERT INTO STOCK_SYMBOL(SYMBOL, COMPANY) VALUES('GOOG', 'Google');
1 row affected (0.009 seconds)
-- 2.3 更新一條數據。結合主鍵使用UPSERT實現。
-- 例如更新 WEB_STAT ，從建表語句可以看到其主鍵爲 HOST、DOMAIN、FEATURE、DATE，
-- 將HOST='EU', DOMAIN='Apple.com', FEATURE='Mac', DATE='2013-01-01 01:01:01.000'的 ACTIVE_VISITOR改爲33
0: jdbc:phoenix:localhost> UPSERT INTO WEB_STAT(HOST, DOMAIN, FEATURE, DATE, ACTIVE_VISITOR) VALUES('EU', 'Apple.com', 'Mac', '2013-01-01 01:01:01.000', 33);
1 row affected (0.191 seconds)
-- 查看結果。發現第一行數據的 ACTIVE_VISITOR 已經更新爲33了
0: jdbc:phoenix:localhost> SELECT * FROM WEB_STAT LIMIT 3;
+-------+-------------+------------+--------------------------+-------+------+-----------------+
| HOST | DOMAIN | FEATURE | DATE | CORE | DB | ACTIVE_VISITOR |
+-------+-------------+------------+--------------------------+-------+------+-----------------+
| EU | Apple.com | Mac | 2013-01-01 01:01:01.000 | 35 | 22 | 33 |
| EU | Apple.com | Store | 2013-01-03 01:01:01.000 | 345 | 722 | 170 |
| EU | Google.com | Analytics | 2013-01-13 08:06:01.000 | 25 | 2 | 6 |
+-------+-------------+------------+--------------------------+-------+------+-----------------+
3 rows selected (0.059 seconds)
-- 3統計不同域名（DOMAIN）下訪客數（ACTIVE_VISITOR）大於100的數據量
0: jdbc:phoenix:localhost> SELECT DOMAIN,COUNT(1) FROM WEB_STAT
. . . . . . . . . . . . .> WHERE ACTIVE_VISITOR>100 GROUP BY DOMAIN;
+-----------------+-----------+
| DOMAIN | COUNT(1) |
+-----------------+-----------+
| Apple.com | 3 |
| Google.com | 1 |
| Salesforce.com | 4 |
+-----------------+-----------+
3 rows selected (0.041 seconds)
-- 4 CASE，將常見的域名轉成中文，其它則返回原始內容
0: jdbc:phoenix:localhost> SELECT (CASE DOMAIN WHEN 'Apple.com' THEN '蘋果公司'
. . . . . . . . . . . . .> WHEN 'Google.com' THEN '谷歌'
. . . . . . . . . . . . .> ELSE DOMAIN END ) AS "域名" FROM WEB_STAT GROUP BY DOMAIN;
+-----------------+
| 域名 |
+-----------------+
| 蘋果公司 |
| 谷歌 |
| Salesforce.com |
+-----------------+
3 rows selected (0.084 seconds)
-- 5 分頁查詢。結合order by ,排序字段最好爲主鍵。LIMIT語法如下
-- [ LIMIT { count } ]
-- [ OFFSET start [ ROW | ROWS ] ]
-- [ FETCH { FIRST | NEXT } [ count ] { ROW | ROWS } ONLY ]
-- 例如接下來，我們對WEB_STAT表的數據按照活躍的訪客數（ACTIVE_VISITOR）倒序進行分頁，
-- 每頁顯示20條數據，從起始的第一條數據開始顯示。
-- LIMIT限制每頁顯示20條數據， OFFSET 爲第一次數據的起始偏移量
0: jdbc:phoenix:localhost> SELECT * FROM WEB_STAT ORDER BY ACTIVE_VISITOR DESC
. . . . . . . . . . . . .> LIMIT 20 OFFSET 0;
+-------+-----------------+------------+--------------------------+-------+-------+-----------------+
| HOST | DOMAIN | FEATURE | DATE | CORE | DB | ACTIVE_VISITOR |
+-------+-----------------+------------+--------------------------+-------+-------+-----------------+
| NA | Salesforce.com | Reports | 2013-01-09 17:36:01.000 | 355 | 432 | 315 |
| NA | Salesforce.com | Login | 2013-01-17 02:20:01.000 | 1235 | 2422 | 243 |
| EU | Google.com | Search | 2013-01-09 01:01:01.000 | 395 | 922 | 190 |
| EU | Apple.com | Store | 2013-01-03 01:01:01.000 | 345 | 722 | 170 |
| NA | Apple.com | Mac | 2013-01-02 04:01:01.000 | 345 | 255 | 155 |
| NA | Salesforce.com | Login | 2013-01-10 01:01:01.000 | 345 | 252 | 150 |
| EU | Salesforce.com | Login | 2013-01-12 01:01:01.000 | 5 | 62 | 150 |
| NA | Apple.com | Login | 2013-01-04 01:01:01.000 | 135 | 2 | 110 |
| NA | Google.com | Search | 2013-01-10 01:05:01.000 | 835 | 282 | 80 |
| NA | Salesforce.com | Login | 2013-01-16 01:01:01.000 | 785 | 782 | 80 |
| NA | Google.com | Analytics | 2013-01-07 06:01:01.000 | 23 | 1 | 57 |
| NA | Google.com | Analytics | 2013-01-14 01:01:01.000 | 65 | 252 | 56 |
| EU | Salesforce.com | Reports | 2013-01-05 04:14:12.000 | 475 | 252 | 53 |
| NA | Google.com | Search | 2013-01-08 08:01:01.000 | 345 | 242 | 46 |
| NA | Salesforce.com | Reports | 2013-01-15 07:09:01.000 | 655 | 426 | 46 |
| NA | Salesforce.com | Login | 2013-01-04 11:01:11.000 | 23 | 56 | 45 |
| NA | Salesforce.com | Dashboard | 2013-01-03 11:01:01.000 | 88 | 66 | 44 |
| EU | Salesforce.com | Dashboard | 2013-01-06 05:04:05.000 | 12 | 22 | 43 |
| EU | Salesforce.com | Reports | 2013-01-02 14:32:01.000 | 125 | 131 | 42 |
| NA | Apple.com | Login | 2013-01-01 01:01:01.000 | 35 | 22 | 40 |
+-------+-----------------+------------+--------------------------+-------+-------+-----------------+
20 rows selected (0.24 seconds)
-- 第二頁，LIMIT限制每頁顯示20條數據， OFFSET 指定爲20表示第二頁的數據的起始偏移量
0: jdbc:phoenix:localhost> SELECT * FROM WEB_STAT ORDER BY ACTIVE_VISITOR DESC
. . . . . . . . . . . . .> LIMIT 20 OFFSET 20;
+-------+-----------------+------------+--------------------------+-------+------+-----------------+
| HOST | DOMAIN | FEATURE | DATE | CORE | DB | ACTIVE_VISITOR |
+-------+-----------------+------------+--------------------------+-------+------+-----------------+
| EU | Apple.com | Mac | 2013-01-01 01:01:01.000 | 35 | 22 | 34 |
| NA | Salesforce.com | Login | 2013-01-17 01:01:01.000 | 355 | 242 | 33 |
| NA | Salesforce.com | Dashboard | 2013-01-11 01:01:01.000 | 335 | 32 | 30 |
| NA | Apple.com | iPad | 2013-01-05 01:01:01.000 | 85 | 2 | 18 |
| NA | Salesforce.com | Reports | 2013-01-09 16:33:01.000 | 35 | 42 | 15 |
| NA | Apple.com | iPad | 2013-01-06 01:01:01.000 | 35 | 22 | 10 |
| NA | Salesforce.com | Login | 2013-01-08 14:11:01.000 | 345 | 242 | 10 |
| NA | Salesforce.com | Login | 2013-01-01 01:01:01.000 | 35 | 42 | 10 |
| NA | Apple.com | Mac | 2013-01-08 01:01:01.000 | 3 | 2 | 10 |
| NA | Salesforce.com | Dashboard | 2013-01-14 04:07:01.000 | 5 | 2 | 9 |
| NA | Apple.com | iPad | 2013-01-07 01:01:01.000 | 9 | 27 | 7 |
| NA | Google.com | Analytics | 2013-01-11 01:02:01.000 | 7 | 2 | 7 |
| NA | Salesforce.com | Reports | 2013-01-15 04:09:01.000 | 65 | 26 | 6 |
| NA | Google.com | Search | 2013-01-12 01:01:01.000 | 8 | 7 | 6 |
| EU | Google.com | Analytics | 2013-01-13 08:06:01.000 | 25 | 2 | 6 |
| EU | Salesforce.com | Reports | 2013-01-13 08:04:04.000 | 355 | 52 | 5 |
| EU | Salesforce.com | Reports | 2013-01-05 03:11:12.000 | 75 | 22 | 3 |
| EU | Salesforce.com | Reports | 2013-01-02 12:02:01.000 | 25 | 11 | 2 |
| NA | Salesforce.com | Login | 2013-01-04 06:01:21.000 | 3 | 52 | 1 |
+-------+-----------------+------------+--------------------------+-------+------+-----------------+
19 rows selected (0.18 seconds)
-- 6 刪除數據。刪除時可以添加條件，也可以模糊匹配刪除
-- 爲了模擬，我們先添加兩條數據
0: jdbc:phoenix:localhost> UPSERT INTO STOCK_SYMBOL VALUES('WAG', 'Walgreens');
1 row affected (0.116 seconds)
0: jdbc:phoenix:localhost> UPSERT INTO STOCK_SYMBOL VALUES('WMT', 'Walmart');
1 row affected (0.008 seconds)
-- 6.1 刪除 SYMBOL='CRM'的一條數據
0: jdbc:phoenix:localhost> DELETE FROM STOCK_SYMBOL WHERE SYMBOL='CRM';
1 row affected (0.089 seconds)
-- 發現已成功刪除
0: jdbc:phoenix:localhost> SELECT * FROM STOCK_SYMBOL;
+---------+-------------+
| SYMBOL | COMPANY |
+---------+-------------+
| AAPL | APPLE Inc. |
| GOOG | Google |
| WAG | Walgreens |
| WMT | Walmart |
+---------+-------------+
4 rows selected (0.022 seconds)
-- 6.2 刪除 COMPANY 爲Wal開頭的所有數據
0: jdbc:phoenix:localhost> DELETE FROM STOCK_SYMBOL WHERE COMPANY LIKE 'Wal%';
2 rows affected (0.034 seconds)
-- 發現有兩條滿足條件的數據已被刪除
0: jdbc:phoenix:localhost> SELECT * FROM STOCK_SYMBOL;
+---------+-------------+
| SYMBOL | COMPANY |
+---------+-------------+
| AAPL | APPLE Inc. |
| GOOG | Google |
+---------+-------------+
2 rows selected (0.017 seconds)
-- 6.3 刪除demo表，如果demo表不是視圖並，對一個的HBase的表和數據也會被刪除。
0: jdbc:phoenix:cdh3> DELETE TABLE "demo";
2.9.1 Join
Phoenix本身提供了Join功能，但比較適用於大表join 小表之類的簡單join，如果要做大表join大表的複雜分析可以藉助Spark來做。
官方文檔介紹中Phoenix支持Join類型如右圖所示有兩種，一種是內連接：INNER JOIN，一種是外連接：OUTER JOIN，其中OUTER JOIN支持LEFT和RIGHT兩種方式。但同時其還支持UNION ALL語法，可以合併來自多個SELECT語句的行數據。
關於這幾種連接文氏圖（Venn diagram）如下所示，INNER JOIN 一般被譯作內連接，內連接查詢能將左表（表 A）和右表（表 B）中能關聯起來的數據連接後返回；LEFT JOIN 一般被譯作左連接，也寫作 LEFT OUTER JOIN ,左連接查詢會返回左表（表 A）中所有記錄，不管右表（表 B）中有沒有關聯的數據,在右表中找到的關聯數據列也會被一起返回；RIGHT JOIN 一般被譯作右連接，也寫作 RIGHT OUTER JOIN ,右連接查詢會返回右表（表 B）中所有記錄，不管左表（表 A）中有沒有關聯的數據,在左表中找到的關聯數據列也會被一起返回。FULL OUTER JOIN 一般被譯作外連接、全連接，實際查詢語句中可以寫作 FULL OUTER JOIN 或 FULL JOIN ,外連接查詢能返回左右表裏的所有記錄，其中左右表裏能關聯起來的記錄被連接後返回。
現在有兩張表，一個是網站數據表websites，主要記錄了網站主鍵id、名稱、連接、網站流量排名、國家等信息。另一張表爲訪問日誌表access_log，主要記錄的有每個網站每日的訪問數量，兩個表通過id來關聯，即websites.id=access_log.site_id ,現在我們通過Phoenix來分別演示INNER JOIN、LEFT [OUTER] JOIN、RIGHT [OUTER] JOIN、FULL [OUTER] JOIN（Phoenix雖然不支持，但可以通過UNION ALL 來實現），對兩個表進行關聯，並根據訪問數量降序排，輸出標識id、名稱、數量、日期。
-- 1 建表
-- 1.1 Websites 表
0: jdbc:phoenix:> CREATE TABLE IF NOT EXISTS TEST."websites" (
. . . . . . . . > "id" UNSIGNED_LONG NOT NULL PRIMARY KEY,
. . . . . . . . > "name" varchar(255) ,
. . . . . . . . > "url" varchar(255) ,
. . . . . . . . > "alexa" UNSIGNED_INT, /*'網站流量排名'*/
. . . . . . . . > "country" varchar(32)
. . . . . . . . > );
No rows affected (0.954 seconds)
-- 1.2 access_log 表
0: jdbc:phoenix:> CREATE TABLE IF NOT EXISTS TEST."access_log" (
. . . . . . . . > "aid" UNSIGNED_LONG NOT NULL,
. . . . . . . . > "site_id" UNSIGNED_LONG NOT NULL,
. . . . . . . . > "count" UNSIGNED_INT,
. . . . . . . . > "access_date" DATE ,
. . . . . . . . > CONSTRAINT PK PRIMARY KEY ("aid", "site_id")
. . . . . . . . > );
No rows affected (0.393 seconds)
-- 1.3 查看創建的表。可以看到本次創建了SCHEM爲TEST的兩張表access_log、websites
0: jdbc:phoenix:> !tables
+------------+--------------+-------------+---------------+----------+------------+----------------------+-----------------+--------------+-----------------+-------------
| TABLE_CAT | TABLE_SCHEM | TABLE_NAME | TABLE_TYPE | REMARKS | TYPE_NAME | SELF_REFERENCING_COL_NAME | REF_GENERATION | INDEX_STATE | IMMUTABLE_ROWS | SALT_BUCKETS | MULTI_TENANT | VIEW_STATEMENT | VIEW_TYPE | INDEX_TYPE | TRANSACTIONAL | IS_NAMESPAC |
+------------+--------------+-------------+---------------+----------+------------+----------------------------+-----------------+--------------+-----------------+--------
| | SYSTEM | CATALOG | SYSTEM TABLE | | | | | | false | null |
| | SYSTEM | FUNCTION | SYSTEM TABLE | | | | | | false | null |
| | SYSTEM | LOG | SYSTEM TABLE | | | | | | true | 32 |
| | SYSTEM | SEQUENCE | SYSTEM TABLE | | | | | | false | null |
| | SYSTEM | STATS | SYSTEM TABLE | | | | | | false | null |
| | | BOOKS | TABLE | | | | | | false | 8 |
| | | STOCK_SYMBOL | TABLE | | | | | | false | null |
| | | WEB_STAT | TABLE | | | | | | false | null |
| | | demo | TABLE | | | | | | false | null |
| | TEST | access_log | TABLE | | | | | | false | null |
| | TEST | websites | TABLE | | | | | | false | null |
+------------+--------------+-------------+---------------+----------+------------+----------------------------+-----------------+--------------+-----------------+--------
-- 2 插入數據。注意：字符串一定用 單引號
-- 2.1 插入 websites 表測試數據
0: jdbc:phoenix:> UPSERT INTO TEST."websites" VALUES(1, 'Google', 'https://www.google.cm/', 1, 'USA');
1 row affected (0.212 seconds)
0: jdbc:phoenix:> UPSERT INTO TEST."websites" VALUES(2, '淘寶', 'https://www.taobao.com/', 4689, 'CN');
1 row affected (0.036 seconds)
0: jdbc:phoenix:> UPSERT INTO TEST."websites" VALUES(3, '豆瓣', 'https://www.douban.com/', 325, 'CN');
1 row affected (0.02 seconds)
0: jdbc:phoenix:> UPSERT INTO TEST."websites" VALUES(4, '微博', 'http://weibo.com/', 20, 'CN');
1 row affected (0.039 seconds)
0: jdbc:phoenix:> UPSERT INTO TEST."websites" VALUES(5, 'Facebook', 'https://www.facebook.com/', 3, 'USA');
1 row affected (0.027 seconds)
0: jdbc:phoenix:> UPSERT INTO TEST."websites" VALUES(7, 'stackoverflow', 'http://stackoverflow.com/', 0, 'IND');
1 row affected (0.016 seconds)
-- 2.2 插入 access_log 表測試數據
0: jdbc:phoenix:> UPSERT INTO TEST."access_log" VALUES(1, 1, 45, '2016-05-10');
1 row affected (0.049 seconds)
0: jdbc:phoenix:> UPSERT INTO TEST."access_log" VALUES(2, 3, 100, '2016-05-13');
1 row affected (0.064 seconds)
0: jdbc:phoenix:> UPSERT INTO TEST."access_log" VALUES(3, 1, 230, '2016-05-14');
1 row affected (0.054 seconds)
0: jdbc:phoenix:> UPSERT INTO TEST."access_log" VALUES(4, 2, 10, '2016-05-14');
1 row affected (0.017 seconds)
0: jdbc:phoenix:> UPSERT INTO TEST."access_log" VALUES(5, 5, 205, '2016-05-14');
1 row affected (0.021 seconds)
0: jdbc:phoenix:> UPSERT INTO TEST."access_log" VALUES(6, 4, 13, '2016-05-15');
1 row affected (0.018 seconds)
0: jdbc:phoenix:> UPSERT INTO TEST."access_log" VALUES(7, 3, 220, '2016-05-15');
1 row affected (0.038 seconds)
0: jdbc:phoenix:> UPSERT INTO TEST."access_log" VALUES(8, 5, 545, '2016-05-16');
1 row affected (0.03 seconds)
0: jdbc:phoenix:> UPSERT INTO TEST."access_log" VALUES(9, 3, 201, '2016-05-17');
1 row affected (0.019 seconds)
0: jdbc:phoenix:> UPSERT INTO TEST."access_log" VALUES(10, 6, 419, '2018-11-21');
1 row affected (0.02 seconds)
-- 3 查看錶中的數據
-- 3.1 查看 websites 表數據
0: jdbc:phoenix:> SELECT * FROM TEST."websites";
+-----+----------------+----------------------------+--------+----------+
| id | name | url | alexa | country |
+-----+----------------+----------------------------+--------+----------+
| 1 | Google | https://www.google.cm/ | 1 | USA |
| 2 | 淘寶 | https://www.taobao.com/ | 4689 | CN |
| 3 | 豆瓣 | https://www.douban.com/ | 325 | CN |
| 4 | 微博 | http://weibo.com/ | 20 | CN |
| 5 | Facebook | https://www.facebook.com/ | 3 | USA |
| 7 | stackoverflow | http://stackoverflow.com/ | 0 | IND |
+-----+----------------+----------------------------+--------+----------+
6 rows selected (0.139 seconds)
-- 3.2 查看 access_log 表數據
0: jdbc:phoenix:> SELECT * FROM TEST."access_log";
+------+----------+--------+--------------------------+
| aid | site_id | count | access_date |
+------+----------+--------+--------------------------+
| 1 | 1 | 45 | 2016-05-10 00:00:00.000 |
| 2 | 3 | 100 | 2016-05-13 00:00:00.000 |
| 3 | 1 | 230 | 2016-05-14 00:00:00.000 |
| 4 | 2 | 10 | 2016-05-14 00:00:00.000 |
| 5 | 5 | 205 | 2016-05-14 00:00:00.000 |
| 6 | 4 | 13 | 2016-05-15 00:00:00.000 |
| 7 | 3 | 220 | 2016-05-15 00:00:00.000 |
| 8 | 5 | 545 | 2016-05-16 00:00:00.000 |
| 9 | 3 | 201 | 2016-05-17 00:00:00.000 |
| 10 | 6 | 419 | 2018-11-21 00:00:00.000 |
+------+----------+--------+--------------------------+
10 rows selected (0.123 seconds)
-- 4 INNER JOIN
-- 可以看到 inner join 返回的是滿足條件的A表和B表都都存在的數據
0: jdbc:phoenix:> SELECT A."name",B."count",B."access_date"
. . . . . . . . > FROM TEST."websites" AS A
. . . . . . . . > INNER JOIN TEST."access_log" AS B
. . . . . . . . > ON A."id"=B."site_id"
. . . . . . . . > ORDER BY B."count" DESC;
+-----------+----------+--------------------------+
| A.name | B.count | B.access_date |
+-----------+----------+--------------------------+
| Facebook | 545 | 2016-05-16 00:00:00.000 |
| Google | 230 | 2016-05-14 00:00:00.000 |
| 豆瓣 | 220 | 2016-05-15 00:00:00.000 |
| Facebook | 205 | 2016-05-14 00:00:00.000 |
| 豆瓣 | 201 | 2016-05-17 00:00:00.000 |
| 豆瓣 | 100 | 2016-05-13 00:00:00.000 |
| Google | 45 | 2016-05-10 00:00:00.000 |
| 微博 | 13 | 2016-05-15 00:00:00.000 |
| 淘寶 | 10 | 2016-05-14 00:00:00.000 |
+-----------+----------+--------------------------+
9 rows selected (0.276 seconds)
-- 5 LEFT JOIN
-- 可以看到 A.name=stackoverflow 在A表存在，但是B表不存在，依然返回作爲結果。
-- 而 B.count=419 的數據A表沒有匹配到，因此就沒有返回。 總之以左表A爲準
-- 注意：這裏和MySQL不一樣的地方是，null值排在最前面（無論是排序是 ASC 還是 DESC）
0: jdbc:phoenix:> SELECT A."name",B."count",B."access_date"
. . . . . . . . > FROM TEST."websites" AS A
. . . . . . . . > LEFT JOIN TEST."access_log" AS B
. . . . . . . . > ON A."id"=B."site_id"
. . . . . . . . > ORDER BY B."count" DESC;
+----------------+----------+--------------------------+
| A.name | B.count | B.access_date |
+----------------+----------+--------------------------+
| stackoverflow | null | |
| Facebook | 545 | 2016-05-16 00:00:00.000 |
| Google | 230 | 2016-05-14 00:00:00.000 |
| 豆瓣 | 220 | 2016-05-15 00:00:00.000 |
| Facebook | 205 | 2016-05-14 00:00:00.000 |
| 豆瓣 | 201 | 2016-05-17 00:00:00.000 |
| 豆瓣 | 100 | 2016-05-13 00:00:00.000 |
| Google | 45 | 2016-05-10 00:00:00.000 |
| 微博 | 13 | 2016-05-15 00:00:00.000 |
| 淘寶 | 10 | 2016-05-14 00:00:00.000 |
+----------------+----------+--------------------------+
10 rows selected (0.169 seconds)
-- 6 RIGHT JOIN
-- 可以看到 B.count=419 在B表存在，但是A表不存在，依然返回作爲結果。
-- 而 A.name=stackoverflow 的數據B表沒有匹配到，因此就沒有返回。 總之以右表B爲準
0: jdbc:phoenix:> SELECT A."name",B."count",B."access_date"
. . . . . . . . > FROM TEST."websites" AS A
. . . . . . . . > RIGHT JOIN TEST."access_log" AS B
. . . . . . . . > ON A."id"=B."site_id"
. . . . . . . . > ORDER BY B."count" DESC;
+-----------+----------+--------------------------+
| A.name | B.count | B.access_date |
+-----------+----------+--------------------------+
| Facebook | 545 | 2016-05-16 00:00:00.000 |
| | 419 | 2018-11-21 00:00:00.000 |
| Google | 230 | 2016-05-14 00:00:00.000 |
| 豆瓣 | 220 | 2016-05-15 00:00:00.000 |
| Facebook | 205 | 2016-05-14 00:00:00.000 |
| 豆瓣 | 201 | 2016-05-17 00:00:00.000 |
| 豆瓣 | 100 | 2016-05-13 00:00:00.000 |
| Google | 45 | 2016-05-10 00:00:00.000 |
| 微博 | 13 | 2016-05-15 00:00:00.000 |
| 淘寶 | 10 | 2016-05-14 00:00:00.000 |
+-----------+----------+--------------------------+
10 rows selected (0.205 seconds)
-- 7 UNION ALL。Phoenix不支持 FULL OUTER JOIN ，但可以通過 UNION ALL 實現
-- FULL OUTER JOIN 等價於 (LEFT OUTER JOIN) UNION ALL (RIGHT OUTER JOIN )
-- 可以看到返回左表和右表都有的所有數據.
-- 注意：目前還不支持 UNION ALL 的左右 SELECT 語句中有 ORDER BY ,否則SQL解析時會報如下錯誤
-- ERROR 602 (42P00): Syntax error. Missing "EOF" at line 5, column 1.
-- 如果要實現排序，可以用子查詢或者視圖來實現
0: jdbc:phoenix:> SELECT A."name",B."count",B."access_date"
. . . . . . . . > FROM TEST."websites" AS A
. . . . . . . . > LEFT JOIN TEST."access_log" AS B
. . . . . . . . > ON A."id"=B."site_id"
. . . . . . . . > UNION ALL
. . . . . . . . > SELECT A."name",B."count",B."access_date"
. . . . . . . . > FROM TEST."websites" AS A
. . . . . . . . > RIGHT JOIN TEST."access_log" AS B
. . . . . . . . > ON A."id"=B."site_id";
+----------------+--------+--------------------------+
| NAME | COUNT | ACCESS_DATE |
+----------------+--------+--------------------------+
| Google | 45 | 2016-05-10 00:00:00.000 |
| Google | 230 | 2016-05-14 00:00:00.000 |
| 淘寶 | 10 | 2016-05-14 00:00:00.000 |
| 豆瓣 | 100 | 2016-05-13 00:00:00.000 |
| 豆瓣 | 220 | 2016-05-15 00:00:00.000 |
| 豆瓣 | 201 | 2016-05-17 00:00:00.000 |
| 微博 | 13 | 2016-05-15 00:00:00.000 |
| Facebook | 205 | 2016-05-14 00:00:00.000 |
| Facebook | 545 | 2016-05-16 00:00:00.000 |
| stackoverflow | null | |
| Google | 45 | 2016-05-10 00:00:00.000 |
| 豆瓣 | 100 | 2016-05-13 00:00:00.000 |
| Google | 230 | 2016-05-14 00:00:00.000 |
| 淘寶 | 10 | 2016-05-14 00:00:00.000 |
| Facebook | 205 | 2016-05-14 00:00:00.000 |
| 微博 | 13 | 2016-05-15 00:00:00.000 |
| 豆瓣 | 220 | 2016-05-15 00:00:00.000 |
| Facebook | 545 | 2016-05-16 00:00:00.000 |
| 豆瓣 | 201 | 2016-05-17 00:00:00.000 |
| | 419 | 2018-11-21 00:00:00.000 |
+----------------+--------+--------------------------+
20 rows selected (0.258 seconds)
-- 7.1 子查詢方式實現 UNION ALL 排序
0: jdbc:phoenix:> SELECT * FROM (
. . . . . . . . > SELECT A."name",B."count",B."access_date"
. . . . . . . . > FROM TEST."websites" AS A
. . . . . . . . > LEFT JOIN TEST."access_log" AS B
. . . . . . . . > ON A."id"=B."site_id"
. . . . . . . . > UNION ALL
. . . . . . . . > SELECT A."name",B."count",B."access_date"
. . . . . . . . > FROM TEST."websites" AS A
. . . . . . . . > RIGHT JOIN TEST."access_log" AS B
. . . . . . . . > ON A."id"=B."site_id"
. . . . . . . . > ) ORDER BY COUNT DESC;
+----------------+--------+--------------------------+
| NAME | COUNT | ACCESS_DATE |
+----------------+--------+--------------------------+
| stackoverflow | null | |
| Facebook | 545 | 2016-05-16 00:00:00.000 |
| Facebook | 545 | 2016-05-16 00:00:00.000 |
| | 419 | 2018-11-21 00:00:00.000 |
| Google | 230 | 2016-05-14 00:00:00.000 |
| Google | 230 | 2016-05-14 00:00:00.000 |
| 豆瓣 | 220 | 2016-05-15 00:00:00.000 |
| 豆瓣 | 220 | 2016-05-15 00:00:00.000 |
| Facebook | 205 | 2016-05-14 00:00:00.000 |
| Facebook | 205 | 2016-05-14 00:00:00.000 |
| 豆瓣 | 201 | 2016-05-17 00:00:00.000 |
| 豆瓣 | 201 | 2016-05-17 00:00:00.000 |
| 豆瓣 | 100 | 2016-05-13 00:00:00.000 |
| 豆瓣 | 100 | 2016-05-13 00:00:00.000 |
| Google | 45 | 2016-05-10 00:00:00.000 |
| Google | 45 | 2016-05-10 00:00:00.000 |
| 微博 | 13 | 2016-05-15 00:00:00.000 |
| 微博 | 13 | 2016-05-15 00:00:00.000 |
| 淘寶 | 10 | 2016-05-14 00:00:00.000 |
| 淘寶 | 10 | 2016-05-14 00:00:00.000 |
+----------------+--------+--------------------------+
20 rows selected (0.291 seconds)
-- 7.2 視圖方式實現 UNION ALL 排序
-- 如果刪除視圖，語法格式爲：DROP VIEW IF EXISTS my_schema.my_view
-- 注意1：目前視圖中還不能包含 AS 關鍵字
-- 注意2：目前子句中不能選擇字段，只能用*號，否則包如下錯誤：
-- ERROR 604 (42P00): Syntax error. Mismatched input. Expecting "ASTERISK", got "TEST" at line 2, column 8.
-- Debug：目前版本存在問題。當前視圖支持的語法比較有限，SELECT中不能有複雜的語法，否則解析時會無法判斷結束標記。可以反饋社區等後期修復
-- CREATE VIEW IF EXISTS TEST."websites_union_access_log" AS
-- SELECT * FROM TEST."websites" LEFT JOIN TEST."access_log"
-- ON TEST."websites"."id"=TEST."access_log"."site_id"
-- UNION ALL
-- SELECT * FROM TEST."websites"
-- RIGHT JOIN TEST."access_log"
-- ON TEST."websites"."id"=TEST."access_log"."site_id";
-- SELECT AB."name",AB."count",AB."access_date" FROM TEST."websites_union_access_log"
-- AS AB ORDER BY AB."count" DESC;
2.10 JDBC
使用JDBC方式連接Phoenix，對數據進行處理
2.10.1 新建一個Maven工程
在IDEA中新建一個Maven項目
2.10.2 pom.xml
<dependencies>
<!-- https://mvnrepository.com/artifact/org.apache.phoenix/phoenix-core -->
<dependency>
<groupId>org.apache.phoenix</groupId>
<artifactId>phoenix-core</artifactId>
<version>5.0.0-HBase-2.0</version>
</dependency>
<!-- https://mvnrepository.com/artifact/org.apache.hadoop/hadoop-common -->
<dependency>
<groupId>org.apache.hadoop</groupId>
<artifactId>hadoop-common</artifactId>
<version>3.0.0</version>
</dependency>
<dependency>
<groupId>junit</groupId>
<artifactId>junit</artifactId>
<version>4.12</version>
</dependency>
</dependencies>
2.10.3 JDBC Client代碼
3 工具
以下兩款常用的工具都支持Win和Mac OS
3.1 DBeaver
在創建Phoenix連接，編輯驅動設那裏點擊 添加文件(F) ，選中前面我們編譯出來的phoenix-5.0.0-HBase-2.0-client.jar包即可。
3.2 SQuirrel SQL
3.2.1 下載
進入這個頁面下載 http://squirrel-sql.sourceforge.net/，這裏可以下載壓縮包版的 Plain zips the latest release for Windows/Linux/MacOS X/others。有三個版本，選在一個版本下載即可，例如這裏下載squirrelsql-3.9.1-optional.zip
3.2.2 解壓和配置
解壓到安裝目錄，然後將phoenix-5.0.0-HBase-2.0-client.jar和phoenix-core-5.0.0-HBase-2.0.jar下載或者軟連接到解壓的包中的lib目錄中。
Su-yuexi:squirrelsql-3.9.1-optional yoreyuan$ cd lib
Su-yuexi:lib yoreyuan$ ln -s ~/.dbeaver-drivers/phoenix-5.0.0-HBase-2.0-client.jar phoenix-5.0.0-HBase-2.0-client.jar
Su-yuexi:lib yoreyuan$ ln -s ~/soft/maven/repository/org/apache/phoenix/phoenix-core/5.0.0-HBase-2.0/phoenix-core-5.0.0-HBase-2.0.jar phoenix-core-5.0.0-HBase-2.0.jar
Su-yuexi:lib yoreyuan$ cd ../
Su-yuexi:squirrelsql-3.9.1-optional yoreyuan$ chmod +x squirrel-sql-mac.sh
3.2.3 啓動
執行啓動腳本，此時會彈出SQuirrel SQL Client Help和SQuirrel SQL Client窗口。如果安裝或其它操作也可以查看幫助文檔。
./squirrel-sql-mac.sh
3.2.4 配置 Phoenix 驅動
- 添加Driver。點擊
Driver - Driver Name: 填寫一個名字
- Example URL:
jdbc:phoenix:cdh2,cdh3:2181 - 在
Java Class Path選在我們前面添加的lib/phoenix-core-5.0.0-HBase-2.0.jar - 點擊
List Drivers，等一會會自動在Class Name：補上org.apache.phoenix.jdbc.PhoenixDriver - OK
- 配置成功後會在左側的
Drivers區域添加上去我們剛配置的Phoenix連接。
3.2.5 連接Phoenix
- 點擊
Aliases中的➕ - Name: 連接的別名
- Driver: 選擇上一步配置好的驅動。URL會自動填入。
- Test。User Name 和 Password 可以不用填寫。
- OK
3.2.6 查詢
執行SQL查詢數據，如上圖右，這裏需要注意的如果我們建表時沒有指定Schema，則創建的表都會在一個空的Schema下（DBeaver工具目前的版本無法顯示空Schema下的表,但是可以正產查詢），點擊點擊SQL，就可以編寫SQL，進行數據的查詢等操作，例如查詢表空schema的表數據：SELECT * FROM WEB_stat;，如果有schema在表前面加上用英文點號分割，例如查詢test下的person表：SELECT * FROM test.person;。
4 資料
這裏推薦一份 雲棲社區的一份資料，可以參考