Sqoop Data Types
When you use Sqoop, some variations apply in the processing. Sqoop supports a subset of data types that database vendors support.
Microsoft SQL Server Data Types
Informatica supports the following Microsoft SQL Server data types when you use Sqoop:
- •Bigint
- •Bit
- •Char
- •Datetime
- •Datetime2
- •Decimal
- •Float
- •Image
- •INT
- •Money
- •Nchar
- •Ntext
- •Numeric
- •Nvarchar
- •Real
- •Smalldatetime
- •Smallint
- •Smallmoney
- •Sysname
- •Timestamp
- •Tinyint
- •Uniqueidentifier
- •Varbinary
- •Varchar
Rules and Guidelines for Sqoop Microsoft SQL Server Data Types
Consider the following rules and guidelines when you configure Microsoft SQL Server data types when you use Sqoop in a streaming mapping:
- •Use the Varbinary data type if the target table that you write to has a column of binary data type because Spark DataFrame does not have a corresponding data type to match its native type and the mapping might throw a runtime exception.
- •Use the Varbinary data type only if you have imported the table while creating the relational target or you selected the Create or replace table at runtime property in the Advance properties. Spark DataFrame does not have a corresponding data type to match its native type.
- •If you export timestamp data, an error occurs as the timestamp data is automatically generated by the database. Use the datetime data type.
Oracle Data Types
Informatica supports the following Oracle data types when you use Sqoop:
- •Blob
- •Char(L)
- •Date
- •Float
- •Long
- •Long Raw
- •Nchar
- •Nclob
- •Number
- •Number(P,S)
- •Nvarchar2
- •Raw
- •Timestamp
- •Varchar
- •Varchar2