Can we split Sqoop job by multiple column combination - hadoop

I am using below Sqoop syntax to split Sqoop job by single column[mostly primary key].
sqoop import --connect jdbc:oracle:thin:#//oracle_server:1521/sid --username xxx --password xxx --table EMPLOYEE --split-by ID -m 10
Can we use multiple columns as combination, if primary key is not available to split Sqoop job?
sqoop import --connect jdbc:oracle:thin:#//oracle_server:1521/sid --username xxx --password xxx --table EMPLOYEE --split-by FIRST_NAME,LAST_NAME -m 10

No. We can use only one column in Split-by otherwise you will get the error message - Invalid column name 'FIRST_NAME,LAST_NAME'.
Please note that if you do not have the split-by column, then you have to explicitly set the --num-mappers 1 or -m 1

Related

incremental load using sqoop from mysql to hive

I am new to sqoop and hive . Please help me with understanding
The count of mysql and hive table are different
mysql is 51 rows (table has primary key and no duplicates ) ad hive is 38rows - first run itself
sqoop job --create mmod -- import --connect "jdbc:mysql://cxln2.c.thelab-240901.internal:3306/retail_db" --username sqoopuser --password-file
/tmp/.mysql-pass.txt --table mod --compression-codec org.apache.hadoop.io.compress.BZip2Codec --hive-import --hive-database encry --hive-table mod2 --h
ive-overwrite --check-column last_update_date --incremental lastmodified --merge-key id --last-value 0 --target-dir /user/user_name/append1sqo
pp
It is not creating target dir in given location , instead it creating in warehouse location
I am trying to schedule a sqoop incremental job , somehow I am doing mistake some where
command : above command
2.1 new rows are added with same date
2.2 delete and update on few rows
Output :
No new updates on given table .
It is not updating lastvalue in sqoop job
How to choose merge-key column in sqoop
Where condition in sqoop
--query "select * from reason where id>20 AND $CONDITIONS"
What is the use of $CONDITIONS and do we need to pass the variable in Linux
Is that possible to track rejected rows in sqoop job

Sqoop export from Hcatalog to MySQL with different col names assign

Now my hive table with columns - id, name
and MySQL table - number, id, name
I want to map id (from hive) with number (from mysql), name (from hive) with id (from mysql).
I use the command :
sqoop export --hcatalog-database <my_db> --hcatalog-table <my_table> --columns "number,id" \
--connect jdbc:mysql://db...:3306/test \
--username <my_user> --password <my_passwd> --table <my_mysql_table>
However, it didn't work.
The same scenario liked this case can work fine [1]. The requirement can be fulfilled by locating the hive table on hdfs and using the following command to achieve.
sqoop export --export-dir /[hdfs_path] --columns "number,id" \
--connect jdbc:mysql://db...:3306/test \
--username <my_user> --password <my_passwd> --table <my_mysql_table>
Is there any solution can fulfill my scenario via Hcatalog?
reference :
[1]. Sqoop export from hive to oracle with different col names, number of columns and order of columns
I didn't used the hcatalog part of sqoop, but as is written in the manual, the next script should do the work:
sqoop export --hcatalog-database <my_db> --hcatalog-table <my_table> --map-column-hive "number,id" \
--connect jdbc:mysql://db...:3306/test \
--username <my_user> --password <my_passwd> --table <my_mysql_table>
This option: --map-column-hive when is used along with --hcatalog, does the work for hcatalog instead of hive.
Hope that this works for you.

error while performing sqoop - merge

I was trying to sqoop merge two data sets by importing the data from the netezza server.
below are the data sets with the numeric as id and letters as name:
Both of the below tables are imported from netezza using the commands:
sqoop import --connect neteeza_url --username uname --password pwd --table sqoop_merge_1 --hive-import --warehouse-dir hdfs_pth --create-hive-table sqoop_merge_1 -m 1
sqoop_merge_1:
1,a
2,b
3,c
4,d
5,e
sqoop_merge_2:
4,z
5,y
and the commands are:
sqoop merge --new-data hdfs_path/sqoop_merge_2 --onto hdfs_path/sqoop_merge_1 --target-dir hdfs_path/sqoop_merge_output --jar-file jar_file_path/sqoop_merge_class_name.jar --class-name sqoop_merge_class_name --merge-key id
I created the jar file by using the codegen command:
sqoop codegen --connect netezza_url --username uname --password -pwd --table sqoop_merge_1
But I am getting the following error:
java.io.IOException: Cannot join values on null key. Did you specify a key column that exists?
Tried all the ways i knew but still getting the error.
Please help.
As you are sure about id column existence, it could be an issue due to case-sensitivity.
Check if you specified ID in Netezza?
If yes, try with --merge-key ID.

Appending Data to hive Table using Sqoop

I am trying to append data to already existing Table in hive.Using the Following command first i import the table from MS-SQL Server to hive.
Sqoop Command:
sqoop import --connect "jdbc:sqlserver://XXX.XX.XX.XX;databaseName=mydatabase" --table "my_table" --where "Batch_Id > 100" --username myuser --password mypassword --hive-import
Now i want to append the data to same existing table in hive where "Batch_Id < 100"
I am using the following Command:
sqoop import --connect "jdbc:sqlserver://XXX.XX.XX.XX;databaseName=mydatabase" --table "my_table" --where "Batch_Id < 100" --username myuser --password mypassword --append --hive-table my_table
This command however runs successfully also updates the HDFS data, but when u connect to hive shell and query the table, the records which are appended are not visible.
Sqoop updated the Data on hdfs "/user/hduser/my_table" but the data on "/user/hive/warehouse/batch_dim" is not updated.
How can reslove this issue.
Regards,
Bhagwant Bhobe
Try using
sqoop import --connect "jdbc:sqlserver://XXX.XX.XX.XX;databaseName=mydatabase"
--table "my_table" --where "Batch_Id < 100"
--username myuser --password mypassword
--hive-import --hive-table my_table
when you are using --hive-import DO NOT use --append parameter.
The Sqoop command you're using (--import) is only for ingesting records into HDFS. You need to use the --hive-import flag to import records into Hive.
See http://sqoop.apache.org/docs/1.4.2/SqoopUserGuide.html#_importing_data_into_hive for more details and for additional import configuration options (you may want to change the document reference to your version of Sqoop, of course).

Configuring Sqoop with Mysql?

I have successfully installed SQOOP now the problem is that how to implement it with RDBMS and how to load data from RDBMS to HDFS using SQOOP.
By Using Sqoop You can Load Data directly to Hive Tables or Store the data in Some target Directory in HDFS
If you Need to copy data from RDBMS into Some directory
sqoop import
--connect ConnectionString
--username username
--password Your_Database_Password {In case no password Do not Specify it}
--table tableName
--col column_name(s) {In case you need to call specific columns}
--target-dir '/tmp/myfolder'
--boundary-query 'Select min,max from table name'
--m 5 {set number of mappers to 5}
--fields-terminated-by ',' {how do you want your data to look in target file}
Boundary Query : This is something you can specify. If you do not specify this , then by default this is run in as an inner query which adds up to a complex query.
If you specify this explicitly then this runs as a normal query and hence the performance is increased.
Also you may want to restrict the number of observation ,say based on column ID, and suppose you need data from ID 1 to 1000. Then using Boundary condition and split-by you will be able to restrict your import data.
--boundary-query "select 0,1000 from employee'
--split-by ID
Split-By : You use Split by on a Sqoop import to specify the column on basis of which split is required. By default,if you do not specify this, sqoop pics up table's primary key as the Split_by column.
Split By picks up data from tables and stores them in different folders based on number of mappers. By Default Number of Mappers are 4.
This may seem unwanted but in case you have a composite primary key or no primary key at all, then sqoop fails to pick up data and may error out.
Note: You may not face any issue if you set the number of mappers to 1. In this case, no split by condition is used since there is only one mapper. So query runs fine. This can be done using
--m 1
If you Need to copy data from RDBMS into Hive Table
sqoop import
--connect ConnectionString
--username username
--password Your_Database_Password {In case no password Do not Specify it}
--table tableName
--boundary-query 'Select min,max from table name'
--m 5 {set number of mappers to 5}
--hive-import
--hive-table serviceorderdb.productinfo
--m 1
Running a query instead of calling entire table itself
sqoop import
--connect ConnectionString
--username username
--password Your_Database_Password
--query 'select name from employees where name like '%s' and $CONDITIONS'
--m 5 {set number of mappers to 5}
--target-dir '/tmp/myfolder'
--fields-terminated-by ',' {how do you want your data to look in target file}
You may see $conditions as extra parameter $CONDITIONS. This is because this time you specified no table and specified a query explicity. When Sqoop runs, it searches for a boundary conditions, which it does not find. Then It Searches for a table and a primary key for applying boundary query which again it will not find. Hence we use $CONDITIONS to explicitly specify that we are not using a query and use default boundry condition from query result.
Checking if your connection is set up properly : For this you can just call list databases and if the you see your data populated then your connection is fine.
$ sqoop list-databases
--connect jdbc:mysql://localhost/
--username root
--password pwd
Connection String for Different Databases :
MYSQL: jdbc:mysql://<hostname>:<port>/<dbname>
jdbc:mysql://127.0.0.1:3306/test_database
Oracle :#//host_name:port_number/service_name
jdbc:oracle:thin:scott/tiger#//myhost:1521/myservicename
You may learn more about sqoop imports from : https://sqoop.apache.org/docs/1.4.1-incubating/SqoopUserGuide.html
By using sqoop import command you can import data from RDBMS to HDFS, Hive and HBase
sqoop import --connect jdbc:mysql://localhost:portnumber/DBName --username root --table emp --password root -m1
By using this command data will be stored in HDFS.
Sample commands to run sqoop import (load data from RDBMS to HDFS):
Postgres
sqoop import --connect jdbc:postgresql://postgresHost/databaseName
--username username --password 123 --table tableName
MySQL
sqoop import --connect jdbc:mysql://mysqlHost/databaseName --username username --password 123 --table tableName
Oracle*
sqoop import --connect jdbc:oracle:thin:#oracleHost:1521/databaseName --username USERNAME --password 123 --table TABLENAME
SQL Server
sqoop import --connect 'jdbc:sqlserver://sqlserverhost:1433;database=dbname;username=<username>;password=<password>' --table tableName
*Sqoop won't find any columns from a table if you don't specify both the username and the table in correct case. Usually, specifying both in uppercase will resolve the issue.
Read the Sqoop User's Guide: https://sqoop.apache.org/docs/1.4.5/SqoopUserGuide.html
I also recommend the Apache Sqoop Cookbook. You will learn how to use import and export tools, do incremental import jobs, save jobs, solve problems with jdbc drivers and much more. http://shop.oreilly.com/product/0636920029519.do

Resources