Migrating from Legacy Exports to Export Pipelines

Widget Exports are deprecated. Existing widget exports keep running for now, but don't set up new ones: use Export Pipelines instead. To move an existing widget export, see Migrating from Legacy Exports to Export Pipelines.

Data Exporters and Widget Exports are being retired and replaced by Export Pipelines. This guide walks you through moving your existing exports, running old and new side by side, and switching over without a gap in your data.

Before you start: Read Getting Started with Export Pipelines for an overview of pipelines, manifests, destinations, and schedules.

Step 1: List Your Current Exports

Write down every legacy export you receive and who uses it:

• Data Exporters: one per channel. Note the channel, the bucket and path, and which file types you use (core_, stats_, conversions_, breakdowns_) at which levels.

• Widget Exports: one per widget. Note the widget, the date range, the schedule, the bucket and path, and the file name.

• Consumers: the scripts, warehouse jobs, or dashboards that read these files.

Your Clarisights team can send you the list of legacy exports set up for your company.

Step 2: Design Your Extracts

In an Export Pipeline, each file you receive is an extract: a list of dimensions, metrics, and optional filters. One pipeline can combine several channels in the same file.

If you use	Create
A widget export	One extract with the widget's dimensions, metrics, and filters. Choose the widget's channels as the pipeline's data sources, and pick a date range preset that matches the widget's date range.
Data Exporter stats_ and core_ files for a level	One extract per level you use, with the time dimension you need (for example Date), the dimensions of that level (for example Campaign), and the metrics you use. Dimensions and metrics come in the same file, so you no longer need to join two files.
Data Exporter conversions_ files	Add the conversion metrics you need to the same extract, or create a separate extract for them.
Data Exporter breakdowns_ files	An extract that adds the breakdown dimension (for example Age or Country) to the level's dimensions, where that dimension is available for your data sources. Ask us if you are not sure.

Only include the columns your consumers actually read. You can add more later with a new manifest version.

Step 3: Prepare the Destination

You can use the same bucket as your legacy exports, but use a different path prefix (for example clarisights/pipelines) so the new files don't mix with the old ones.

Check the permissions. Export Pipelines need s3:PutObject and s3:PutObjectAcl on S3, or storage.objects.create (the Storage Object Creator role) on GCS. If your legacy exports already write to the bucket, these are usually in place. See Configuring Export Pipeline Destinations & Permissions for all options.

Step 4: Choose the Schedule

Decide when each export should run and which dates it should cover:

Legacy behaviour	Export Pipeline equivalent
Data Exporter daily run	A daily cron (for example 0 8 * * *) with the same number of days, such as last_30d
Data Exporter extra run for the last 3 days	A second schedule with a different cron and today_3d
Widget export at a chosen hour, every N hours	A cron at that hour, repeating every N hours (for example 0 */6 * * * for every 6 hours)

Set the schedule's timezone to the one your consumers expect. Legacy exports use your company's timezone.

Step 5: Ask Us to Create the Pipeline

Send the details from steps 2 to 4 to your Clarisights team through the messenger support on the platform. We create the pipeline, its manifest, the destination, and the schedule, and confirm once the first run has delivered files. If you need history in the new format, we can also run a one-time export of past dates.

Step 6: Update Your Consumers

The new files are laid out differently. Update the scripts or jobs that read them:

Item	Legacy exports	Export Pipelines
Path	from_clarisights/<company>/<channel>/v<version>/... (Data Exporters) or a fixed path and file name (Widget Exports)	{path_prefix}/{pipeline_name}/v{manifest_version}/{extract_id}/{epoch_time}_{start_date}_{end_date}.csv.gz by default, adjustable with the file name template
Files per run	Several files per channel and level, possibly split into parts	One data file per extract, never split, plus a .meta.json file
Column names	Fixed per channel	Dimension names as shown in Clarisights, and metric keys. The .meta.json file maps each column to its name in Clarisights.
Export time column	created_at (Data Exporters)	_extracted_at
Compression	gzip, or as configured for widget exports	gzip by default, or none

Step 7: Run Both Side by Side

Keep the legacy export running while the new pipeline delivers files. For a week or two, compare the two for the same dates and timezone, for example total spend per day and per channel. Small differences can come from rounding. Larger ones usually come from a different date range, timezone, or filter.

Step 8: Switch Off the Legacy Export

When your consumers read only the new files and the numbers match, tell us which legacy exports to switch off. Files that were already delivered stay in your bucket.

In case of any queries, feel free to reach out to us from the messenger support on the platform.