Create a complex file data object to read data from or write data to HDFS or View File System (ViewFS).
Select a project or folder in the
Object Explorer
view.
Click
File
New
Data Object
.
Select
Complex File Data Object
and click
Next
.
The
New Complex File Data Object
dialog box appears.
Optionally, enter a name for the data object.
Click
Browse
next to the
Location
option and select the target project or folder.
In the
Resource Format
list, select any of the following formats:
Intelligent Structure Model: to read any format that an intelligent structure parses.
Binary: to read any resource format.
Avro: to read an Avro resource.
Parquet: to read a Parquet resource.
JSON: to read a JSON resource.
Orc: to read an Orc resource.
XML: to read an XML resource.
Intelligent structure model is supported only in Spark mode.
In the
Access Type
list, select
Connection
or
File
.
Select
Connection
to access a file on HDFS or ViewFS. Click
Browse
next to the
Connection
option and select an HDFS or ViewFS connection. Click
Add
next to the
Selected Resource
option to add a resource to the data object. If a default Metadata Access Service is not set, a message appears to configure the Metadata Access Service. Click
OK
and set one Metadata Access Service as default. After you set a default Metadata Access Service, the
Add Resource
dialog box appears. If the Metadata Access Service does not exist, contact the Informatica administrator to create a new Metadata Access Service in the domain. Navigate or search for the resources to add to the data object and click
OK
.
Select
File
to access a file on your local system. Click
Browse
next to the
Resource Location
option and select the file that you want to add. Click
Fetch
. The selected file is added to the
Selected Resources
list.
To use an intelligent structure model, for the
Selected Resource
option, browse to and select the appropriate
.amodel
file.
From the
Available OS Profiles
list, select an operating system profile. You can use the
Available OS Profiles
to increase security and to isolate the design-time user environment when you import and preview metadata from a Hadoop cluster.
The Developer tool displays the
Available OS Profiles
list only if the Metadata Access Service is enabled to use operating system profiles. The Metadata Access Service imports the metadata with the default operating system profile assigned to the user. You can change the operating system profile from the list of available operating system profiles.
Click
Finish
.
The data object appears under the Physical Data Objects category in the project or folder in the
Object Explorer
view. A read and write operation is created for the data object. Depending on whether you want to use the complex file data object as a source or target, you can edit the read or write operation properties. You can also create multiple read and write operations for a complex file data object. For a data object with an intelligent structure model, create a read operation. You cannot use a write transformation for a data object with an intelligent structure model in a mapping.
The complex file data object write operation goes through and the mapping runs successfully even if you have unconnected ports for required fields in the Parquet resource type. The NULL values are inserted in the target object when such a mapping runs. The complex file data object read operation results in an error while reading NULL values from the Parquet resource as Parquet Example Object Model does not support NULL read.
For a read operation with an intelligent structure model, specify the path to the input file. In the