Hive 中的複合資料結構簡介以及一些函式的用法說明
目前 hive 支援的複合資料型別有以下幾種:
map (key1, value1, key2, value2, ...) Creates a map with the given key/value pairs struct (val1, val2, val3, ...) Creates a struct with the given field values. Struct field names will be col1, col2, ... named_struct (name1, val1, name2, val2, ...) Creates a struct with the given field names and values. (as of Hive 0.8.0) array (val1, val2, ...) Creates an array with the given elements create_union (tag, val1, val2, ...) Creates a union type with the value that is being pointed to by the tag parameter
一、map、struct、array 這3種的用法:
1、Array的使用
建立資料庫表,以array作為資料型別 create table person(name string,work_locations array<string>) ROW FORMAT DELIMITED FIELDS TERMINATED BY 't' COLLECTION ITEMS TERMINATED BY ','; 資料 biansutao beijing,shanghai,tianjin,hangzhou linan changchu,chengdu,wuhan 入庫資料 LOAD DATA LOCAL INPATH '/home/hadoop/person.txt' OVERWRITE INTO TABLE person; 查詢 hive> select * from person; biansutao ["beijing","shanghai","tianjin","hangzhou"] linan ["changchu","chengdu","wuhan"] Time taken: 0.355 seconds hive> select name from person; linan biansutao Time taken: 12.397 seconds hive> select work_locations[0] from person; changchu beijing Time taken: 13.214 seconds hive> select work_locations from person; ["changchu","chengdu","wuhan"] ["beijing","shanghai","tianjin","hangzhou"] Time taken: 13.755 seconds hive> select work_locations[3] from person; NULL hangzhou Time taken: 12.722 seconds hive> select work_locations[4] from person; NULL NULL Time taken: 15.958 seconds
2、Map 的使用
建立資料庫表 create table score(name string, score map<string,int>) ROW FORMAT DELIMITED FIELDS TERMINATED BY 't' COLLECTION ITEMS TERMINATED BY ',' MAP KEYS TERMINATED BY ':'; 要入庫的資料 biansutao '數學':80,'語文':89,'英語':95 jobs '語文':60,'數學':80,'英語':99 入庫資料 LOAD DATA LOCAL INPATH '/home/hadoop/score.txt' OVERWRITE INTO TABLE score; 查詢 hive> select * from score; biansutao {"數學":80,"語文":89,"英語":95} jobs {"語文":60,"數學":80,"英語":99} Time taken: 0.665 seconds hive> select name from score; jobs biansutao Time taken: 19.778 seconds hive> select t.score from score t; {"語文":60,"數學":80,"英語":99} {"數學":80,"語文":89,"英語":95} Time taken: 19.353 seconds hive> select t.score['語文'] from score t; 60 89 Time taken: 13.054 seconds hive> select t.score['英語'] from score t; 99 95 Time taken: 13.769 seconds
3、Struct 的使用
建立資料表
CREATE TABLE test(id int,course struct<course:string,score:int>)
ROW FORMAT DELIMITED
FIELDS TERMINATED BY 't'
COLLECTION ITEMS TERMINATED BY ',';
資料
1 english,80
2 math,89
3 chinese,95
入庫
LOAD DATA LOCAL INPATH '/home/hadoop/test.txt' OVERWRITE INTO TABLE test;
查詢
hive> select * from test;
OK
1 {"course":"english","score":80}
2 {"course":"math","score":89}
3 {"course":"chinese","score":95}
Time taken: 0.275 seconds
hive> select course from test;
{"course":"english","score":80}
{"course":"math","score":89}
{"course":"chinese","score":95}
Time taken: 44.968 seconds
select t.course.course from test t;
english
math
chinese
Time taken: 15.827 seconds
hive> select t.course.score from test t;
80
89
95
Time taken: 13.235 seconds
4、資料組合 (不支援組合的複雜資料型別)
LOAD DATA LOCAL INPATH '/home/hadoop/test.txt' OVERWRITE INTO TABLE test;
create table test1(id int,a MAP<STRING,ARRAY<STRING>>)
row format delimited fields terminated by 't'
collection items terminated by ','
MAP KEYS TERMINATED BY ':';
1 english:80,90,70
2 math:89,78,86
3 chinese:99,100,82
LOAD DATA LOCAL INPATH '/home/hadoop/test1.txt' OVERWRITE INTO TABLE test1;
二、hive中的一些不常見函式的用法:
常見的函式就不廢話了,和標準sql類似,下面我們要聊到的基本是HQL裡面專有的函式,
hive裡面的函式大致分為如下幾種:Built-in、Misc.、UDF、UDTF、UDAF
我們就挑幾個標準SQL裡沒有,但是在HIVE SQL在做統計分析常用到的來說吧。
1、array_contains (Collection Functions)
這是內建的對集合進行操作的函式,用法舉例:
create EXTERNAL table IF NOT EXISTS userInfo (id int,sex string, age int, name string, email string,sd string, ed string) ROW FORMAT DELIMITED FIELDS TERMINATED BY 't' location '/hive/dw';
select * from userinfo where sex='male' and (id!=1 and id !=2 and id!=3 and id!=4 and id!=5) and age < 30;
select * from (select * from userinfo where sex='male' and !array_contains(split('1,2,3,4,5',','),cast(id as string))) tb1 where tb1.age < 30;
其中建表所用的測試資料你可以用如下連結的指令碼自動生成:
http://my.oschina.net/leejun2005/blog/76631
2、get_json_object (Misc. Functions)
測試資料:
first {"store":{"fruit":[{"weight":8,"type":"apple"},{"weight":9,"type":"pear"}],"bicycle":{"price":19.951,"color":"red1"}},"email":"amy@only_for_json_udf_test.net","owner":"amy1"} third first {"store":{"fruit":[{"weight":9,"type":"apple"},{"weight":91,"type":"pear"}],"bicycle":{"price":19.952,"color":"red2"}},"email":"amy@only_for_json_udf_test.net","owner":"amy2"} third first {"store":{"fruit":[{"weight":10,"type":"apple"},{"weight":911,"type":"pear"}],"bicycle":{"price":19.953,"color":"red3"}},"email":"amy@only_for_json_udf_test.net","owner":"amy3"} third
create external table if not exists t_json(f1 string, f2 string, f3 string) row format delimited fields TERMINATED BY ' ' location '/test/json'
select get_json_object(t_json.f2, '$.owner') from t_json;
SELECT * from t_json where get_json_object(t_json.f2, '$.store.fruit[0].weight') = 9;
SELECT get_json_object(t_json.f2, '$.non_exist_key') FROM t_json;
這裡尤其要注意UDTF的問題,官方文件有說明:
json_tuple A new json_tuple() UDTF is introduced in hive 0.7. It takes a set of names (keys) and a JSON string, and returns a tuple of values using one function. This is much more efficient than calling GET_JSON_OBJECT to retrieve more than one key from a single JSON string. In any case where a single JSON string would be parsed more than once, your query will be more efficient if you parse it once, which is what JSON_TUPLE is for. As JSON_TUPLE is a UDTF, you will need to use the LATERAL VIEW syntax in order to achieve the same goal. For example,
select a.timestamp, get_json_object(a.appevents, '$.eventid'), get_json_object(a.appenvets, '$.eventname') from log a;
should be changed to
select a.timestamp, b.*
from log a lateral view json_tuple(a.appevent, 'eventid', 'eventname') b as f1, f2;
UDTF(User-Defined Table-Generating Functions) 用來解決 輸入一行輸出多行(On-to-many maping) 的需求。
通過Lateral view可以方便的將UDTF得到的行轉列的結果集合在一起提供服務,因為直接在SELECT使用UDTF會存在限制,即僅僅能包含單個欄位,不光是多個UDTF,僅僅單個UDTF加上其他欄位也是不可以,hive提示在UDTF中僅僅能有單一的表示式。如下: hive> select my_test(“abcef:aa”) as qq,’abcd’ from sunwg01; FAILED: Error in semantic analysis: Only a single expression in the SELECT clause is supported with UDTF’s
使用Lateral view可以實現上面的需求,Lateral view語法如下: lateralView: LATERAL VIEW udtf(expression) tableAlias AS columnAlias (‘,’ columnAlias)* fromClause: FROM baseTable (lateralView)* hive> create table sunwg ( a array, b array ) > ROW FORMAT DELIMITED > FIELDS TERMINATED BY ‘t’ > COLLECTION ITEMS TERMINATED BY ‘,’; OK Time taken: 1.145 seconds hive> load data local inpath ‘/home/hjl/sunwg/sunwg.txt’ overwrite into table sunwg; Copying data from file:/home/hjl/sunwg/sunwg.txt Loading data to table sunwg OK Time taken: 0.162 seconds hive> select * from sunwg; OK [10,11] ["tom","mary"] [20,21] ["kate","tim"] Time taken: 0.069 seconds hive> > SELECT a, name > FROM sunwg LATERAL VIEW explode(b) r1 AS name; OK [10,11] tom [10,11] mary [20,21] kate [20,21] tim Time taken: 8.497 seconds hive> SELECT id, name > FROM sunwg LATERAL VIEW explode(a) r1 AS id > LATERAL VIEW explode(b) r2 AS name; OK 10 tom 10 mary 11 tom 11 mary 20 kate 20 tim 21 kate 21 tim Time taken: 9.687 seconds
3、parse_url_tuple
測試資料:
url1 http://facebook.com/path1/p.php?k1=v1&k2=v2#Ref1 url2 https://cwiki.apache.org/confluence/display/Hive/LanguageManual+UDF#LanguageManualUDF-getjsonobject url3 https://www.google.com.hk/#hl=zh-CN&newwindow=1&safe=strict&q=hive+translate+example&oq=hive+translate+example&gs_l=serp.3...10174.11861.6.12051.8.8.0.0.0.0.132.883.0j7.7.0...0.0...1c.1j4.8.serp.0B9C1T_n0Hs&bav=on.2,or.&bvm=bv.44770516,d.aGc&fp=e13e41a6b9dab3f6&biw=1241&bih=589
create external table if not exists t_url(f1 string, f2 string) row format delimited fields TERMINATED BY ' ' location '/test/url';
SELECT f1, b.* FROM t_url LATERAL VIEW parse_url_tuple(f2, 'HOST', 'PATH', 'QUERY', 'QUERY:k1') b as host, path, query, query_id;
結果:
url1 facebook.com /path1/p.php k1=v1&k2=v2 v1 url2 cwiki.apache.org /confluence/display/Hive/LanguageManual+UDF NULL NULL url3 www.google.com.hk / NULL NULL
4、explode
explode 是一個 hive 內建的表生成函式:Built-in Table-Generating Functions (UDTF),主要是解決 1 to N 的問題,即它可以把一行輸入拆成多行,比如一個 array 的每個元素拆成一行,作為一個虛表輸出。它有如下需要注意的地方:
Using the syntax "SELECT udtf(col) AS colAlias..." has a few limitations:
No other expressions are allowed in SELECT
SELECT pageid, explode(adid_list) AS myCol... is not supported
UDTF's can't be nested
SELECT explode(explode(adid_list)) AS myCol... is not supported
GROUP BY / CLUSTER BY / DISTRIBUTE BY / SORT BY is not supported
SELECT explode(adid_list) AS myCol ... GROUP BY myCol is not supported
從上面的原理與語法上可知,
- select 列中不能 udtf 和其它非 udtf 列混用,
- udtf 不能巢狀,
- 不支援 GROUP BY / CLUSTER BY / DISTRIBUTE BY / SORT BY
- 還有 select 中出現的 udtf 一定需要列別名,否則會報錯:
SELECT explode(myCol) AS myNewCol FROM myTable;
SELECT explode(myMap) AS (myMapKey, myMapValue) FROM myMapTable;
SELECT posexplode(myCol) AS pos, myNewCol FROM myTable;
5、lateral view
lateral view 是Hive中提供給UDTF的conjunction,它可以解決UDTF不能新增額外的select列的問題。當我們想對hive表中某一列進行split之後,想對其轉換成1 to N的模式,即一行轉多列。hive不允許我們在UDTF函式之外,再新增其它select語句。
如下,我們想將登入某個遊戲的使用者id放在一個欄位user_ids裡,對每一行資料用UDTF後輸出多行。
select game_id, explode(split(user_ids,'\[\[\[')) as user_id from login_game_log where dt='2014-05-15' ;
FAILED: Error in semantic analysis: UDTF's are not supported outside the SELECT clause, nor nested in expressions。
提示語法分析錯誤,UDTF不支援函式之外的select 語句,如果我們想支援怎麼辦呢?接下來就是Lateral View 登場的時候了。
Lateral view 其實就是用來和像類似explode這種UDTF函式聯用的。lateral view 會將UDTF生成的結果放到一個虛擬表中,然後這個虛擬表(1 to N)會和輸入行即每個game_id進行join 來達到連線UDTF外的select欄位的目的(源表和拆分的虛表按行做行內 1 join N 的直接連線),這也是為什麼 LATERAL VIEW udtf(expression) 後面需要表別名和列別名的原因。
Lateral View Syntax
lateralView: LATERAL VIEW udtf(expression) tableAlias AS columnAlias (',' columnAlias)*
fromClause: FROM baseTable (lateralView)*
可以看出,可以在2個地方用Lateral view:
- 在udtf前面用
- 在from baseTable後面用
例如:
pageid adid_list
front_page [1, 2, 3]
contact_page [3, 4, 5]
SELECT pageid, adid
FROM pageAds LATERAL VIEW explode(adid_list) adTable AS adid;
pageid adid
front_page 1
front_page 2
front_page 3
contact_page 3
contact_page 4
contact_page 5
From語句後可以跟多個Lateral View。
A FROM clause can have multiple LATERAL VIEW clauses. Subsequent LATERAL VIEWS can reference columns from any of the tables appearing to the left of the LATERAL VIEW.
給定資料:
Array<int> col1 Array<string> col2
[1, 2] [a", "b", "c"]
[3, 4] [d", "e", "f"]
轉換目標:
想同時把第一列和第二列拆開,類似做笛卡爾乘積。
我們可以這樣寫:
SELECT myCol1, myCol2 FROM baseTable
LATERAL VIEW explode(col1) myTable1 AS myCol1
LATERAL VIEW explode(col2) myTable2 AS myCol2;
還有一種情況,如果UDTF轉換的Array是空的怎麼辦呢?
在Hive0.12裡面會支援outer關鍵字,如果UDTF的結果是空,預設會被忽略輸出。
如果加上outer關鍵字,則會像left outer join 一樣,還是會輸出select出的列,而UDTF的輸出結果是NULL。
總結:
- Lateral View通常和UDTF一起出現,為了解決UDTF不允許在select欄位的問題。
- Multiple Lateral View可以實現類似笛卡爾乘積。
- Outer關鍵字可以把不輸出的UDTF的空結果,輸出成NULL,防止丟失資料。
三、ref:
http://blog.csdn.net/wf1982/article/details/7474601 http://www.cnblogs.com/ggjucheng/archive/2013/01/08/2850797.html http://www.oratea.net/?p=650 https://cwiki.apache.org/confluence/display/Hive/LanguageManual+UDF#LanguageManualUDF-parseurltuple https://cwiki.apache.org/confluence/display/Hive/Tutorial
http://blog.csdn.net/inte_sleeper/article/details/7196114 hive lateral view語句:列拆分成行
https://cwiki.apache.org/confluence/display/Hive/LanguageManual+UDF#LanguageManualUDF-explode
http://blog.csdn.net/oopsoom/article/details/26001307 Lateral View用法 與 Hive UDTF explode
http://bit.ly/2bDuVxS 助力大資料的複雜統計分析-Hive視窗函式