Troubleshoot an import or update that is slow, stuck or failed

This page applies to customer sites: the organisation data portals on Team Cloud, Business Cloud, Enterprise plans. It covers what you see in the Data Manager when an import or update doesn’t finish, what common errors mean, and a checklist for when a scheduled update doesn’t run.

What the status on a dataset means

On the Datasets tab of the Data Manager, a dataset shows a status next to its name:

Status	What it means
Importing	The first import of a new dataset is running. A progress percentage is shown.
Updating	An update to an existing dataset is running.
Publishing	The data is in and Koordinates is publishing it.
Draft	A draft exists that hasn’t been published.
Import error	The first import failed. The dataset has no published data yet.
Draft error	An update failed. The published version is unchanged.
Publish error	The data imported but publishing failed.
Update Available	A connected source has newer data than the dataset.
Scheduled	The dataset has a scheduled update.

The time an import takes depends on the size and type of the data, and on how much else your site is processing. Large files and many-file imports take longer. Progress that moves, even slowly, means the import is still working.

If an import or update looks stuck

Refresh the Data Manager and check the status

Reload the page, then check the dataset’s status and progress percentage. If the percentage has changed since you last looked, wait.
Check whether it has failed instead

A failed import shows Import error or Draft error. Open the dataset to see a banner such as Error during import, Error during update or Error during publish. Koordinates also emails the person who started the import when it fails. The email names the dataset and, for common problems, says what’s wrong. See Errors you might see.
Check the source data

Check that the files open in desktop GIS software, that a shapefile includes all its parts, and that a connected source is still reachable. If you’re importing from a connected source, scan it again and check the Logs tab for errors. For connection problems, see Troubleshoot a data source connection.
Cancel the import, fix the data and try again

While a dataset is importing or updating, the Cancel import button appears beside its status. Cancelling stops the import immediately and can’t be undone. If it’s the dataset’s first import, cancelling also deletes the dataset. Cancelling an update stops it immediately too.
Delete an errored draft

An errored draft can block further updates. Open the dataset and delete the errored draft: the banner offers Delete errored draft. If the first import failed, the banner offers Delete dataset, so you can start again with different source data.
Contact support if it still doesn’t finish

Raise a ticket with Koordinates support. Give the dataset’s name and ID (the number in its address), the time you started the import, the status you can see, and the text of any error email. Don’t send passwords or API keys.

Errors you might see

When an import fails for a known reason, the email to the person who started it says why. The reasons Koordinates reports are:

Message	What to do
The files you selected don’t share the same attributes	When you import several files into one dataset, they must have the same attributes (columns). Import them separately, or fix the columns first.
There are invalid geometries in the input data	Repair the geometries in your GIS software and import again.
The data contains more than one type of geometry	A dataset can hold only one geometry type, such as polygons. Split the data by geometry type and import each part separately.
Some data lies outside the known extent of the coordinate system	The projection may be wrong, or some coordinates are wrong. Check the coordinate system of the source and fix it.
The primary key you specified was not unique, or contained null values	Choose a different primary key, or correct the values. See Set a primary key for a data layer.
There were no valid features found	Check that the file has data and that its geometry and projection are readable.
You selected more than one type of data	You can’t merge vector, raster, grid and non-spatial data into one dataset. Import each type separately.

For any other error, the email says that support staff have been notified. It may include the address of a support ticket. If you don’t hear back, raise a ticket yourself.

Why didn’t my layer update automatically?

A dataset updates on a schedule only when all of the following are true. Work down the list.

A schedule is switched on for the dataset

Open the dataset in the Data Manager and select the Scheduled Updates tab. Enable must be switched on. A dataset with a schedule shows Scheduled beside its name. If the tab isn’t there, see step 2.
Your plan and dataset support scheduled updates

The Scheduled Updates tab is shown only if your plan includes scheduled updates and the dataset isn’t a set or a Kart repository dataset. The tab says the dataset isn’t compatible if it isn’t connected to a single external data source. Scheduled updates are for vector and table datasets, so raster, grid and point cloud datasets can’t have one. Uploaded files can’t be used for scheduled updates: the data must come from a connected data source. On the Team Cloud plan, the number of datasets with a schedule is limited to 10, and saving a schedule fails once you reach the limit.
The schedule is what you expect

You can schedule a dataset daily, weekly or monthly, at a time on the hour, in a time zone you choose. If a schedule was set another way, such as through the API, and is too complex for the page, the page shows Complex schedule currently configured. Select Clear existing schedule to set a new one.
The source has newer data

At each scheduled time, Koordinates checks the connected source for changes. If the source has no new, changed or deleted data since the dataset was last updated, nothing happens and no new version is made. Check the source’s data and its scan history under the Logs tab. For a PostgreSQL table, Koordinates detects a change from the table’s columns and the insert, update and delete counts that PostgreSQL keeps for it.
The source is still connected and its data is still there

If the source item the dataset was made from has been deleted, scheduled updates stop. If the item is gone for a long time, Koordinates gives up trying until someone updates the dataset from a live source. Check the source on the Sources tab, and update the dataset manually from the source if needed.
The new data matches the dataset

A scheduled update only runs on its own if the new data is close enough to the existing dataset that nobody needs to check it. It doesn’t run if:
- the projection has changed,
- the geometry type has changed,
- a column’s data type has changed, or the data type of the source has changed (for example, vector to raster), or
- the columns have changed (added or removed) and Continue scheduled update if dataset schema has changed is off. Columns you omitted when importing don’t count.
Fix the source or update the dataset manually, then review the result. To let scheduled updates continue when columns are added or removed, switch on Continue scheduled update if dataset schema has changed.
The previous import isn’t stuck or errored

A scheduled update is skipped while the latest version of the dataset is importing or has errored. Delete the errored draft, or switch on Continue to update if an errored draft exists so that Koordinates replaces an errored draft and carries on. Then save the schedule.
The update itself didn’t fail

If the scheduled update ran but failed, the dataset shows Draft error. See If an import or update looks stuck.

You can scan a source or update a dataset by hand at any time without waiting for the schedule. To set or change a schedule, see Schedule regular updates of your data. You can also manage schedules with the API: see Admin API recipes.