Configure a dataflow to ingest batch data from a cloud storage source in the UI

This tutorial provides steps on how to configure a dataflow to bring batch data from your cloud storage source to Adobe Experience Platform.

Getting started

NOTE
In order to create a dataflow to bring batch data from a cloud storage, you must already have access to an authenticated cloud storage source. If you do not have access, go to the sources overview for a list of cloud storage sources that you can create an account with.

This tutorial requires a working understanding of the following components of Experience Platform:

Supported file formats

Cloud storage sources for batch data supports the following file formats for ingestion:

  • Delimiter-separated values (DSV): Any single-character value can be used as a delimiter for DSV-formatted data files.
  • JavaScript Object Notation (JSON): JSON-formatted data files must be XDM-compliant.
  • Apache Parquet: Parquet-formatted data files must be XDM-compliant.
  • Compressed files: JSON and delimited files can be compressed as: bzip2, gzip, deflate, zipDeflate, tarGzip, and tar.

Add data

After creating your cloud storage account, the Add data step appears, providing an interface for you to explore your cloud storage file hierarchy and select the folder or specific file that you want to bring to Platform.

  • The left part of the interface is a directory browser, displaying your cloud storage file hierarchy.
  • The right part of the interface lets you preview up to 100 rows of data from a compatible folder or file.

Select the root folder to access your folder hierarchy. From here, you can select a single folder to ingest all files in the folder recursively. When ingesting an entire folder, you must ensure that all files in that folder share the same data format and schema.

Once you have selected a folder, the right interface updates to a preview of the contents and structure of the first file in the selected folder.

During this step, you can make several configurations to your data, before proceeding. First, select Data format and then select the appropriate data format for your file in the dropdown panel that appears.

The following table displays the appropriate data formats for the supported file types:

File type
Data format
CSV
Delimited
JSON
JSON
Parquet
XDM Parquet

Select a column delimiter

After configuring your data format, you can set a column delimiter when ingesting delimited files. Select the Delimiter option and then select a delimiter from the dropdown menu. The menu displays the most frequently used options for delimiters, including a comma (,), a tab (\t), and a pipe (|).

If you prefer to use a custom delimiter, select Custom and enter a single-character delimiter of your choice in the pop up input bar.

Ingest compressed files

You can also ingest compressed JSON or delimited files by specifying their compression type.

In the Select data step, select a compressed file for ingestion and then select its appropriate file type and whether it’s XDM-compliant or not. Next, select Compression type and then select the appropriate compressed file type for your source data.

To bring a specific file to Platform, select a folder, and then select the file that you want to ingest. During this step, you can also preview file contents of other files within a given folder by using the preview icon beside a file name.

When finished, select Next.

Provide dataflow details

The Dataflow detail page allows you to select whether you want to use an existing dataset or a new dataset. During this process, you can also configure your data to be ingested to Profile, and enable settings like Error diagnostics, Partial ingestion, and Alerts.

Use an existing dataset

To ingest data into an existing dataset, select Existing dataset. You can either retrieve an existing dataset using the Advanced search option or by scrolling through the list of existing datasets in the dropdown menu. Once you have selected a dataset, provide a name and a description for your dataflow.

Use a new dataset

To ingest into a new dataset, select New dataset and then provide an output dataset name and an optional description. Next, select a schema to map to using the Advanced search option or by scrolling through the list of existing schemas in the dropdown menu. Once you have selected a schema, provide a name and a description for your dataflow.

Enable Profile and error diagnostics

Next, select the Profile dataset toggle to enable your dataset for Profile. This allows you to create a holistic view of an entity’s attributes and behaviors. Data from all Profile-enabled datasets will be included in Profile and changes are applied when you save your dataflow.

Error diagnostics enables detailed error message generation for any erroneous records that occur in your dataflow, while Partial ingestion allows you to ingest data containing errors, up to a certain threshold that you manually define. See the partial batch ingestion overview for more information.

Enable alerts

You can enable alerts to receive notifications on the status of your dataflow. Select an alert from the list to subscribe to receive notifications on the status of your dataflow. For more information on alerts, see the guide on subscribing to sources alerts using the UI.

When you are finished providing details to your dataflow, select Next.

Map data fields to an XDM schema

The Mapping step appears, providing you with an interface to map the source fields from your source schema to their appropriate target XDM fields in the target schema.

Platform provides intelligent recommendations for auto-mapped fields based on the target schema or dataset that you selected. You can manually adjust mapping rules to suit your use cases. Based on your needs, you can choose to map fields directly, or use data prep functions to transform source data to derive computed or calculated values. For comprehensive steps on using the mapper interface and calculated fields, see the Data Prep UI guide.

Once your source data is successfully mapped, select Next.

Schedule ingestion runs

IMPORTANT
It is strongly recommended to schedule your dataflow for one-time ingestion when using the FTP source.

The Scheduling step appears, allowing you to configure an ingestion schedule to automatically ingest the selected source data using the configured mappings. By default, scheduling is set to Once. To adjust your ingestion frequency, select Frequency and then select an option from the dropdown menu.

TIP
Interval and backfill are not visible during a one-time ingestion.

scheduling

If you set your ingestion frequency to Minute, Hour, Day, or Week, then you must set an interval to establish a set time frame between every ingestion. For example, an ingestion frequency set to Day and an interval set to 15 means that your dataflow is scheduled to ingest data every 15 days.

During this step, you can also enable backfill and define a column for the incremental ingestion of data. Backfill is used to ingest historical data, while the column you define for incremental ingestion allows new data to be differentiated from existing data.

See the table below for more information on scheduling configurations.

Field
Description
Frequency
The frequency in which an ingestion happens. Selectable frequencies include Once, Minute, Hour, Day, and Week.
Interval
An integer that sets the interval for the selected frequency. The interval’s value should be a non-zero integer and should be set to greater than or equal to 15.
Start time
A UTC timestamp indicating when the very first ingestion is set to occur. Start time must be greater than or equal to your current UTC time.
Backfill
A boolean value that determines what data is initially ingested. If backfill is enabled, all current files in the specified path will be ingested during the first scheduled ingestion. If backfill is disabled, only the files that are loaded in between the first run of ingestion and the start time will be ingested. Files loaded prior to start time will not be ingested.
NOTE
For batch ingestion, every ensuing dataflow selects files to be ingested from your source based on their last modified timestamp. This means that batch dataflows select files from the source that are either new or have been modified since the last flow run. Furthermore, you must ensure that there’s a sufficient time span between file upload and a scheduled flow run because files that are not entirely uploaded to your cloud storage account before the scheduled flow run time may not be picked up for ingestion.

When finished configuring your ingestion schedule, select Next.

Review your dataflow

The Review step appears, allowing you to review your new dataflow before it is created. Details are grouped within the following categories:

  • Connection: Shows the source type, the relevant path of the chosen source file, and the amount of columns within that source file.
  • Assign dataset & map fields: Shows which dataset the source data is being ingested into, including the schema that the dataset adheres to.
  • Scheduling: Shows the active period, frequency, and interval of the ingestion schedule.

Once you have reviewed your dataflow, click Finish and allow some time for the dataflow to be created.

Next steps

By following this tutorial, you have successfully created a dataflow to bring in data from an external cloud storage, and gained insight on monitoring datasets. To learn more about creating dataflows, you can supplement your learning by watching the video below. Additionally, incoming data can now be used by downstream Platform services such as Real-Time Customer Profile and Data Science Workspace. See the following documents for more details:

WARNING
The Platform UI shown in the following video is out-of-date. Please refer to the documentation above for the latest UI screenshots and functionality.
Transcript
Hi, there. I’m going to give you a quick overview of how to ingest data from your Cloud Storage sources into Adobe Experience Platform. Data ingestion is a fundamental step to getting your data in Experience Platform so you can use it to build 360 degree Real-time Customer Profiles and use them to provide meaningful experiences. You can ingest data from a wide variety of sources, such as Adobe applications Cloud-based storage, databases, and many others. When you log into Platform you will see sources in the left navigation. Clicking Sources will take you to Source Catalog screen. Where we can see all of the source connectors currently available in Platform. Note that there are source connectors for Adobe applications CRM solutions, Cloud storage providers and more. Let’s explore how to ingest data from Cloud Storage into Experience Platform. Each source has its specific configuration details but the general configuration for Cloud Storage sources are somewhat similar. For our video, let’s use the Amazon S3 Cloud Storage. Select the desired source. When selecting up a source connector for the first time you will be provided with an option to configure. Since this is our first time creating an S3 account let’s click on Creating a New Account and provide the source connection details. Complete the required fields for account authentication and then initiate a source connection request. If the connection is successful, click Next to proceed to Data Selection. In this step, we choose the source file for data ingestion and verify the file data format. Note that the ingested file data can be formatted as JSON, XDM parquet or delimited. Currently, for delimited files you have an option to preview sample data of the source file. Let’s proceed to the next step to assign a target Dataset for the incoming data. Let’s choose the New Dataset option and provide a Dataset name and description. To create a Dataset, you need to have an associated schema. Using the schema finder, assign a schema to this particular Dataset. Upon selecting a schema for this Dataset Experience Platform performs a mapping within the source file field and the target field. This mapping is performed based on the title and type of the fields. This pre-mapping of standard fields are editable. You can add a new field mapping to map a source field to a target field. Add Calculated Field option lets you run functions on source fields to prepare the data for ingestion. For example, we can combine the first name and the last name field into a calculated field using the Concatenation function before ingesting the data into a Dataset field. We can also preview the sample result of a calculated field. After reviewing the field mapping, we can also preview data to see how the ingested data will get stored in your Dataset. The mapping looks good, let’s move to the next step. Scheduling lets you choose a frequency at which data should flow from source to a Dataset. Let’s select a frequency of 15 minutes for this video and set a start time for data flow. To let historical data to be ingested enable the backfill option. Backfill is a Boolean value that determines what data is initially ingested. If backfill is enabled, all current files in the specific pot will be ingested during the first scheduled ingestion. If backfill is disabled, only the files that are loaded in between the first run of ingestion and the start time will be ingested. Let’s move to the data flow step, provide a name for your data flow. In the data flow detail step, the Partial Ingestion toggle allows you to enable or disable the use of partial batch ingestion. The Error Threshold allows you to set the percentage of acceptable errors before they enter batch fields. By default, this value is set to 5%. Let’s review the source configuration details and save your changes. We do not see any data flow run statuses as we have set a frequency of 15 minutes for our data flow runs. So let’s wait for the data flow run to trigger, let’s refresh the page and you can now see that our data flow run status has been completed. Open the data flow run to view more details about the activity. Our last data flow run was completed successfully with a few failed records. Do you wonder why the most recent data flow run was successful and when it had failed records. That’s because we enabled partial ingestion when we set up the data flow and chosen error threshold of 5%. Since we enabled error diagnosis for our data flows, we can also see the error code and description in the Data Flow Run Overview window. Experience Platform lets users preview or download the error diagnosis to determine what went wrong with the failed records. Let’s go back to the Data Flow Activity tab. At this point, we verified that the data flow was completed successfully from the source to our Dataset. Let’s open our Dataset to verify the data flow and its activities. You can open the Luma Customer Loyalty Dataset right from the data flow window or you can access it to using the Datasets option from the left navigation. Under the Dataset Activity we can see a quick summary of ingested batches and failed batches during a specific time window. Scroll down to view the ingested batch ID. Each batch represents actual data ingestion from a source connector to a target Dataset. With Real-time Customer Profile we can see each customer’s holistic view that combines data from multiple channels, including online, offline, CRM and then third-party data. To enable our Dataset for a Real-time Profile, ensure that the associated schema is enabled for Real-time Profile. Once a schema is enabled for profile it cannot be disabled or deleted. Also, fields cannot be removed from the schema after this point. These implications are essential to keep in mind when working with the data in your production environment. It is recommended to verify and test the data ingestion process to capture and address any issues that may arise before enabling the Dataset and schema for the profile. Now let’s enable profile for our Dataset and save our changes. In the next successful batch run data ingested into our Dataset will be used for creating Real-time Customer Profiles. You can also preview the ingested data into the Dataset using the Preview Dataset option. Finally, we can also add any governance label to the Dataset fields to restrict its usage under the Data Governance tab.
A final confirmation step is to inspect a profile that should have been updated in the last ingestion. Look on Profiles in the left navigation and then click on Browse. Find a profile by using an identity namespace and a corresponding identity value. Look up a profile using one of its identities. In this case, let’s use the loyalty ID and make sure it has the new fields. Adobe Experience Platform allows data to be ingested from external sources while providing you with the ability to structure, label labor and enhance incoming data using Platform services. -

Appendix

The following sections provide additional information for working with source connectors.

Monitor your dataflow

Once your dataflow has been created, you can monitor the data that is being ingested through it to view information on ingestion rates, success, and errors. For more information on how to monitor dataflow, visit the tutorial on monitoring accounts and dataflows in the UI.

Update your dataflow

To update configurations for your dataflows scheduling, mapping, and general information, visit the tutorial on updating sources dataflows in the UI

Delete your dataflow

You can delete dataflows that are no longer necessary or were incorrectly created using the Delete function available in the Dataflows workspace. For more information on how to delete dataflows, visit the tutorial on deleting dataflows in the UI.

recommendation-more-help
337b99bb-92fb-42ae-b6b7-c7042161d089