Collect structured datasets, not just individual answers.
Collect large CSV or Excel returns within a digital service. Stream rows through background validation, explain errors, and let respondents review and confirm the dataset before it reaches your operational systems.
Validate up to five million data cells across up to 200 configured columns, without loading the entire dataset into the browser. Availability depends on your account and release.
Collect datasets instead of individual answers
When a return contains many records, ask for a structured dataset within the service. Respondents can upload CSV or Excel data instead of entering every record as an individual answer.
Bulk data collection brings dataset upload, validation and confirmation into the service journey. It is separate from Data Grids, which support direct tabular entry.
Define the expected structure
Define up to 200 dataset columns with text, integer, decimal, date, boolean or fixed-choice types. Apply required values, lengths, ranges and uniqueness rules to check the information your service needs.
Validate up to five million data cells, counted using the configured columns. For example, a 10-column return can contain up to 500,000 data rows within this limit.
Upload CSV or Excel data
Accept CSV or XLSX within the service, with worksheet selection for Excel. CSV rows stream directly; Excel input is staged to temporary disk and worksheet rows are read as a stream.
Background processing, visible progress, a preview and error reports keep respondents informed without opening the entire dataset in their browser.
Set the input-file maximum explicitly: 10 MB is the default and 20 GB is the platform ceiling. Store policies, configured security processing and service capacity can impose a lower limit.
Plan the size and processing path
File size, dataset size and security processing have separate limits. Choose the path that fits your return and the destination’s policies. MB/GB are decimal; MiB is the binary API allowance.
| Capability | Published limit | What to plan |
|---|---|---|
| Input file | 20 GB ceiling; 10 MB default. | Set an explicit component limit. Store policies and processing capacity can reduce it. |
| Dataset validation | Five million data cells; up to 200 columns. | For example, 500,000 data rows across 10 configured columns. |
| Browser correction | Up to 1,000 rows per selection/edit batch. | Correct supported values in manageable batches before confirmation. |
| Optional external-store Govform AV | Current scanned-transfer ceiling: 5 GB. | Applies to scan-first and parallel modes; lower configured limits can apply. |
| Optional CDR | 50 MB default or a lower configured cap. | Single-file components only; always scan first. Confirm the reconstructed output suits dataset ingestion. |
| Base64 API action body | 25 MiB combined raw files per action. | Encoding adds about one third. Stream larger artifacts instead. |
Validate and explain errors
Check uploaded data against the defined column types and validation rules. Error reports help respondents understand which information needs attention.
Keep progress visible while processing runs, then use the validation results to guide the next step.
Correct supported values
Let respondents correct supported values in a spreadsheet-style browser view where configured. Work in selections and edit batches of up to 1,000 rows, then validate and confirm the resulting return.
Keep the source file as separate evidence while the validated CSV incorporates confirmed corrections. Availability depends on your account and release.
Review and confirm the return
Give respondents a chance to review their uploaded dataset and validation results before confirming. Explicit confirmation keeps the decision to submit clear within the journey.
Uploading a file and confirming a return are distinct steps. Make the review and confirmation stage part of the service you design.
Preserve versions and source evidence
Retain the source upload separately from the confirmed CSV, so authorised teams can understand both the evidence supplied and the corrected dataset accepted.
Where Response history is retained, select current confirmed data, the first submitted revision or a specific submitted revision for download or onward delivery. Access and retention settings control availability.
Use our storage or yours, then stream data onward
Choose Govform-managed storage or your own SharePoint, AWS S3, Google Cloud Storage or Azure Blob Storage. Keep QA and Production connections separate and retain your own store’s access and lifecycle controls.
Incoming uploads to your own supported store can use Govform scan-first, parallel scanning or no Govform AV when your destination provides the agreed controls. Govform storage uses configured in-place scanning. CDR always scans first, with separate processing limits.
Deliver the confirmed CSV or selected source evidence using Upload files actions, or stream it through the authenticated Bulk Data API. These paths avoid the 25 MiB Base64 API-body allowance; provider, network and execution limits still apply.
- Govform-managed storage
- SharePoint
- AWS S3
- Google Cloud Storage
- Azure Blob Storage
Bulk uploads compared with Data Grids
Choose the input approach that fits the respondent task. Bulk uploads collect a prepared dataset; Data Grids let respondents enter tabular information directly within a service.
Optional browser correction of an uploaded dataset does not make bulk data collection the same capability as a Data Grid.
| Task | Bulk data collection | Data Grids |
|---|---|---|
| How information arrives | Structured CSV or Excel upload, with worksheet selection for Excel. | Direct tabular entry, including spreadsheet paste and repeated rows. |
| Working with values | Validation rules, error reports and optional correction of supported values. | Copying, calculations, conditional visibility and column validation. |
| Completing the task | Background processing, progress, review and explicit confirmation. | Tabular entry as part of the service journey. |
| Submitted evidence | Versioned bulk returns and retained source files. | A direct-entry capability; do not treat it as a bulk file upload. |
For services that collect many structured records
Provider and regulatory returns: collect a prepared dataset using the columns and rules required by the service.
Grant monitoring and organisational reporting: gather structured records alongside the wider questions in a digital journey.
Recurring reporting: use bulk data collection where a Collection needs a dataset return, while Collection programmes coordinate expected respondents, dates and review.
Frequently asked questions
Make structured returns part of your digital service
Bring your dataset structure, expected file size, storage destination and scanning requirements. We can help you choose the upload, validation, correction and delivery path for your deployment.