Re-importing a dataset
Updating your dataset is a seamless process. You can work on a new, draft version of your data, without affecting the existing published version.
Often, the original data which feeds into your dataset will change over time. For example, if your dataset represents your customer base, then the underlying data will change as you gain customers or learn more about them.
When this happens, you can update your dataset. Updating your dataset replaces the existing data with new data, but doesn't change anything else about its configuration. So any queries you've written will continue to work, but will reference the latest data.
Updating your dataset is a seamless process. You can work on a new, draft version of your data, without affecting the existing published version. While you're working on the draft version, any queries will still reference the published version. Once you're ready, you can go ahead and publish your draft, replacing the existing data.
To update your dataset:
- Starting from the Platform, use the Datasets tab to open the dataset's Bunker, as before.
- Import and normalize your data - again, following the same process as you did previously.
- This time, when you have normalized your data and come to the Publish page, you'll see something like the following.
On the second row, you'll see your existing, published dataset. At this stage, queries are still referencing this version.
On the first row you'll see your new draft. You can compare the two versions in detail, including the detailed breakdown of the representation fill rates, to confirm that you are ready to publish your draft.
- If you're happy, click Publish. Your draft becomes the new published version, replacing the previous published version, and all queries now reference the new data.
- If you're not happy, return to the Normalize page to work further on your data.
Once you have published your draft, you cannot revert to the previous version. Be sure you are ready before clicking Publish.