First commit

.gitattributes

@@ -0,0 +1,2 @@
+*.csv filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text

CH07 SourceCode/CH07 SourceCode/clickstream_code.txt

@@ -0,0 +1,29 @@
+CREATE DATABASE clickstream;
+
+USE clickstream;
+
+CREATE TABLE wikilogs (
+  	previous_id 	STRING,
+  	current_id 	STRING,
+  	no_occurences	INT,
+  	previous_title 	STRING,
+  	current_title  	STRING,
+  	type		STRING)
+ROW FORMAT DELIMITED
+FIELDS TERMINATED BY '09'
+STORED AS textfile;
+
+LOAD DATA INPATH '/tmp/wikiclickstream/2015_01_clickstream.tsv.gz' OVERWRITE INTO TABLE wikilogs;
+
+CREATE VIEW wikilogs_view (no_occurences, previous_title, current_title)
+AS SELECT no_occurences, previous_title, current_title FROM wikilogs;
+
+SELECT * FROM wikilogs_view
+SORT BY no_occurences DESC;
+
+
+SELECT * FROM wikilogs_view
+WHERE previous_title = 'other-facebook'
+SORT BY no_occurences DESC;
+
+

CH07 SourceCode/CH07 SourceCode/json1

@@ -0,0 +1 @@
+ {"_id":"5774091c3227dc593d06aff2","index":0,"guid":"3ce0d275-317d-4936-aedb-2a7844b6d54d","isActive":false,"balance":"$1,912.20","picture":"http://placehold.it/32x32","age":27,"eyeColor":"brown","name":"Fry Schmidt","gender":"male","company":"ZORK","email":"[email protected]","phone":"+1 (956) 510-2736","address":"158 Tompkins Avenue, Rosedale, New York, 2134","about":"Et mollit Lorem id sunt dolor in id aliquip nisi incididunt in elit cupidatat exercitation. Minim nulla fugiat cillum cillum exercitation velit veniam. Lorem excepteur et esse adipisicing pariatur ipsum ut excepteur. Quis adipisicing ullamco esse velit in eiusmod veniam eiusmod Lorem. Elit proident ex velit culpa excepteur deserunt laboris duis irure ipsum voluptate nostrud cupidatat. Consequat eiusmod culpa laborum culpa nulla anim et enim sunt voluptate.\r\n","registered":"2015-05-16T05:55:04 +05:00","latitude":65.239382,"longitude":31.355514,"tags":["ea","consectetur","reprehenderit","Lorem","ea","nisi","et"],"friends":[{"id":0,"name":"Gomez Richmond"},{"id":1,"name":"Deanne Mcintyre"},{"id":2,"name":"Best Dyer"}],"greeting":"Hello, Fry Schmidt! You have 8 unread messages.","favoriteFruit":"apple"}

CH07 SourceCode/CH07 SourceCode/json2

@@ -0,0 +1 @@
+{"_id":"577424548ab78dc174ac0dae","index":0,"guid":"83659acc-ddb7-4194-a1d5-cd9eee5a76b5","isActive":false,"balance":"$2,041.83","picture":"http://placehold.it/32x32","age":36,"eyeColor":"brown","name":{"first":"Roseann","last":"Perry"},"company":"RAMJOB","email":"[email protected]","phone":"+1 (959) 463-3581","address":"203 Pleasant Place, Rehrersburg, Mississippi, 4612","about":"Cupidatat laboris commodo excepteur in incididunt cupidatat sit minim non laborum proident nisi. Occaecat sit eu ea irure occaecat aute amet velit do labore enim sit irure nostrud. Ad ullamco minim tempor esse. Aliquip sit culpa amet sint ea minim in qui magna minim Lorem in voluptate. Ea nulla eu nulla sunt ex. Nulla ea magna cillum esse velit amet voluptate.","registered":"Saturday, April 9, 2016 12:55 PM","latitude":"-11.748526","longitude":"160.379242","tags":["pariatur","laboris","ipsum","sint","eiusmod"],"range":[0,1,2,3,4,5,6,7,8,9],"friends":[{"id":0,"name":"Jessica Lowe"},{"id":1,"name":"Lessie Hoffman"},{"id":2,"name":"Madeline Arnold"}],"greeting":"Hello, Roseann! You have 7 unread messages.","favoriteFruit":"banana"}

CH07 SourceCode/CH07 SourceCode/json3

@@ -0,0 +1 @@
+{"_id":"5774245438f862f0b8121f41","index":5,"guid":"580ff472-9036-40b2-aa3c-9085f305d6b4","isActive":false,"balance":"$2,252.98","picture":"http://placehold.it/32x32","age":36,"eyeColor":"brown","name":{"first":"Lori","last":"Pacheco"},"company":"LUDAK","email":"[email protected]","phone":"+1 (891) 415-2253","address":"290 Rochester Avenue, Cannondale, Guam, 7856","about":"Qui fugiat nostrud qui laborum Lorem excepteur. Minim exercitation esse mollit irure fugiat eiusmod proident sit Lorem incididunt. Dolor ex ipsum tempor est eu duis exercitation. Enim ea ullamco mollit proident labore eiusmod excepteur magna Lorem anim.","registered":"Tuesday, February 10, 2015 8:07 AM","latitude":"75.805649","longitude":"138.091539","tags":["ullamco","in","voluptate","reprehenderit","sunt"],"range":[0,1,2,3,4,5,6,7,8,9],"friends":[{"id":0,"name":"Byrd Meyers"},{"id":1,"name":"Weeks Miles"},{"id":2,"name":"Marquez Pace"}],"greeting":"Hello, Lori! You have 6 unread messages.","favoriteFruit":"banana"}

CH07 SourceCode/CH07 SourceCode/json_code.txt

@@ -0,0 +1,72 @@
+CREATE TABLE json_table (
+json string);
+
+LOAD DATA INPATH '/tmp/json_data/json1' INTO TABLE json_table;
+
+SELECT get_json_object(json_table.jcol, '$') FROM json_table; 
+
+select get_json_object(json_table.json, ‘$.balance) as balance,
+             get_json_object(json_table, ‘$.gender) as gender,
+             get_json_object(json_table.json, ‘$.phone) as phone,
+             get_json_object(json.table.json, ‘$.friends.name) as friendname
+from json_table;
+
+ADD JAR /usr/local/Hive-JSON-Serde/json-serde/target/json-serde-1.3.8-SNAPSHOT-jar-with-dependencies.jar;
+
+CREATE TABLE json_serde_table (
+  id string,
+  about string,
+  address string,
+  age int,
+  balance string,
+  company string,
+  email string,
+  eyecolor string,
+  favoritefruit string,
+  friends array<struct<id:int, name:string>>,
+  gender string,
+  greeting string,
+  guid string,
+  index int,
+  isactive boolean,
+  latitude double,
+  longitude double,
+  name string,
+  phone string,
+  picture string,
+  registered string,
+  tags array<string>)
+ROW FORMAT SERDE 'org.openx.data.jsonserde.JsonSerDe'
+WITH SERDEPROPERTIES ( "mapping._id" = "id" )
+
+LOAD DATA INPATH '/tmp/json_data/json1' INTO TABLE json_serde_table;
+
+SELECT address, friends.name FROM json_serde_table;
+
+CREATE TABLE json_serde_table (
+  id string,
+  about string,
+  address string,
+  age int,
+  balance string,
+  company string,
+  email string,
+  eyecolor string,
+  favoritefruit string,
+  friends array<struct<id:int, name:string>>,
+  gender string,
+  greeting string,
+  guid string,
+  index int,
+  isactive boolean,
+  latitude double,
+  longitude double,
+  name string,
+  phone string,
+  picture string,
+  registered string,
+  tags array<string>)
+ROW FORMAT SERDE 'org.apache.hive.hcatalog.data.JsonSerDe'
+STORED AS TEXTFILE;
+
+

ESM Ch 5 and 8 ExampleData/Script_EqualJoin.txt

@@ -0,0 +1,38 @@
+USE census;
+CREATE TABLE census.personname (
+  persid         int,
+  firstname      string,
+  lastname       string
+)
+CLUSTERED BY (persid) INTO 1 BUCKETS
+STORED AS orc 
+TBLPROPERTIES('transactional' = 'true');
+INSERT INTO TABLE census.personname 
+VALUES 
+(0,'Albert','Ape'),
+(1,'Bob','Burger'),
+(2,'Charlie','Clown'),
+(3,'Danny','Drywer');
+
+
+CREATE TABLE census.address (
+  persid         int,
+  postname      string
+)
+CLUSTERED BY (persid) INTO 1 BUCKETS
+STORED AS orc 
+TBLPROPERTIES('transactional' = 'true');
+INSERT INTO TABLE census.address 
+VALUES 
+(1,'KA13'),
+(2,'KA9'),
+(10,'SW1');
+
+SELECT personname.firstname, 
+personname.lastname,
+address.postname 
+FROM
+census.personname
+JOIN
+census.address
+ON (personname.persid = address.persid);

ESM Ch 5 and 8 ExampleData/Script_MapJoin.txt

@@ -0,0 +1,9 @@
+USE census;
+
+SELECT /*+ MAPJOIN(b) */ personname.firstname, 
+personname.lastname
+FROM
+census.personname
+LEFT SEMI JOIN
+census.address
+ON (personname.persid = address.persid);

ESM Ch 5 and 8 ExampleData/Script_MultiJoin.txt

@@ -0,0 +1,26 @@
+USE census;
+
+CREATE TABLE census.account (
+  persid       int,
+  bamount      int
+)
+CLUSTERED BY (persid) INTO 1 BUCKETS
+STORED AS orc 
+TBLPROPERTIES('transactional' = 'true');
+INSERT INTO TABLE census.account 
+VALUES 
+(1,12),
+(2,9);
+
+SELECT personname.firstname, 
+personname.lastname,
+address.postname,
+account.bamount
+FROM
+census.personname
+JOIN
+census.address
+ON (personname.persid = address.persid)
+JOIN
+census.account
+ON (personname.persid = account.persid);

ESM Ch 5 and 8 ExampleData/Script_OuterJoin.txt

@@ -0,0 +1,30 @@
+USE census;
+
+SELECT personname.firstname, 
+personname.lastname,
+address.postname 
+FROM
+census.personname
+LEFT JOIN
+census.address
+ON (personname.persid = address.persid);
+
+
+SELECT personname.firstname, 
+personname.lastname,
+address.postname 
+FROM
+census.personname
+RIGHT JOIN
+census.address
+ON (personname.persid = address.persid);
+
+
+SELECT personname.firstname, 
+personname.lastname,
+address.postname 
+FROM
+census.personname
+FULL OUTER JOIN
+census.address
+ON (personname.persid = address.persid);

ESM Ch 5 and 8 ExampleData/Script_PersonDelete.txt

@@ -0,0 +1,15 @@
+USE census;
+CREATE TABLE census.person30 (
+  persid         int,
+  lastname       string,
+  firstname      string
+)
+CLUSTERED BY (persid) INTO 1 BUCKETS
+STORED AS orc 
+TBLPROPERTIES('transactional' = 'true');
+INSERT INTO TABLE census.person30 VALUES (0,'A','B'),(2,'X','Y');
+SELECT * FROM census.person30;
+
+DELETE FROM census.person30 WHERE persid = 0;
+
+SELECT * FROM census.person30;

ESM Ch 5 and 8 ExampleData/Script_PersonDirectory.txt

@@ -0,0 +1,11 @@
+hadoop fs -ls
+hadoop fs -mkdir 'exampleoutput'
+hive
+
+USE census;
+INSERT OVERWRITE DIRECTORY 'exampleoutput' ROW FORMAT  DELIMITED  FIELDS TERMINATED BY ',' 
+SELECT persid, firstname, lastname FROM person;
+exit;
+
+hadoop fs -ls 'exampleoutput'
+hadoop fs -cat 'exampleoutput/000000_0'

ESM Ch 5 and 8 ExampleData/Script_PersonHub.txt

@@ -0,0 +1,9 @@
+USE census;
+CREATE TABLE personhub (
+  persid         int
+);
+INSERT OVERWRITE TABLE personhub SELECT DISTINCT persid FROM Person;
+SELECT persid FROM personhub where persid < 11;
+INSERT INTO TABLE personhub SELECT DISTINCT persid FROM Person;
+SELECT persid FROM personhub where persid < 11;
+

ESM Ch 5 and 8 ExampleData/Script_PersonLike.txt

@@ -0,0 +1,6 @@
+USE census;
+CREATE TABLE person40 LIKE person;
+SELECT * FROM person40;
+
+INSERT INTO TABLE person40 VALUES (0,'Bob','Burger'),(1,'Charlie','Clown');
+SELECT * FROM person40;

ESM Ch 5 and 8 ExampleData/Script_PersonTable.txt

@@ -0,0 +1,11 @@
+CREATE DATABASE census;
+USE census;
+CREATE TABLE person (
+  persid         int,
+  lastname       string,
+  firstname      string
+)
+ROW FORMAT  DELIMITED  FIELDS TERMINATED BY ',';
+
+LOAD DATA LOCAL INPATH 'file:///root/hive/example/person001' OVERWRITE INTO TABLE person;
+SELECT persid, lastname, firstname FROM person;

ESM Ch 5 and 8 ExampleData/Script_PersonUpdate.txt

@@ -0,0 +1,16 @@
+USE census;
+CREATE TABLE census.person20 (
+  persid         int,
+  lastname       string,
+  firstname      string
+)
+CLUSTERED BY (persid) INTO 1 BUCKETS
+STORED AS orc 
+TBLPROPERTIES('transactional' = 'true');
+INSERT INTO TABLE person20 VALUES (0,'A','B'),(2,'X','Y');
+
+SELECT * FROM census.person20;
+
+UPDATE census.person20 SET lastname = 'SS' WHERE persid = 0;
+
+SELECT * FROM census.person20 WHERE persid = 0;

ESM Ch 5 and 8 ExampleData/Script_PersonValues.txt

@@ -0,0 +1,3 @@
+USE census;
+INSERT INTO TABLE personhub VALUES (0);
+SELECT persid FROM personhub WHERE persid = 0;

ESM Ch 5 and 8 ExampleData/Script_SemiJoin.txt

@@ -0,0 +1,9 @@
+USE census;
+
+SELECT personname.firstname, 
+personname.lastname
+FROM
+census.personname
+LEFT SEMI JOIN
+census.address
+ON (personname.persid = address.persid);

ESM Ch 5 and 8 ExampleData/datawarehouse/01retrieve/MapR_Retrieve001.txt

@@ -0,0 +1,43 @@
+--#######################################################################################################################################
+--#######################################################################################################################################
+--#######################################################################################################################################
+DROP DATABASE retrievedb CASCADE;
+
+--#######################################################################################################################################
+CREATE DATABASE IF NOT EXISTS retrievedb;
+USE retrievedb;
+
+--#######################################################################################################################################
+--DROP TABLE retrievedb.rawfirstname;
+--DROP TABLE retrievedb.rawlastname;
+--DROP TABLE retrievedb.rawpersonname;
+--#######################################################################################################################################
+CREATE TABLE IF NOT EXISTS retrievedb.rawfirstname (
+  firstnameid    string,
+  firstname      string,
+  sex            string
+)
+ROW FORMAT DELIMITED FIELDS TERMINATED BY ',';
+
+LOAD DATA LOCAL INPATH 'file:///home/mapr/ExampleData/00rawdata/rawfirstname.csv' OVERWRITE INTO TABLE retrievedb.rawfirstname;
+
+--#######################################################################################################################################
+CREATE TABLE IF NOT EXISTS retrievedb.rawlastname (
+  lastnameid    string,
+  lastname      string
+)
+ROW FORMAT DELIMITED FIELDS TERMINATED BY ',';
+
+LOAD DATA LOCAL INPATH 'file:///home/mapr/ExampleData/00rawdata/rawlastname.csv' OVERWRITE INTO TABLE retrievedb.rawlastname;
+
+--#######################################################################################################################################
+CREATE TABLE IF NOT EXISTS retrievedb.rawperson (
+  persid         string,
+  firstnameid    string,
+  lastnameid     string
+)
+ROW FORMAT DELIMITED FIELDS TERMINATED BY ',';
+
+LOAD DATA LOCAL INPATH 'file:///home/mapr/ExampleData/00rawdata/rawperson.csv' OVERWRITE INTO TABLE retrievedb.rawperson;
+
+--#######################################################################################################################################