Business colleagues discussing paperwork around an office table
Illustrative stock photograph. Pexels. Not a product screenshot or evidence of a test.

Power Automate Get Files Returns Only 100: Check Scope, Top Count and Pagination

A Get files action returning 100 results may be following its default retrieval limit rather than reaching the end of the library. Inspect scope, filters, Top Count and pagination before treating that output as a complete inventory.

Use a read-only test flow first. Do not connect an unverified retrieval step directly to bulk deletion, approval or external notifications. This guide concerns the SharePoint Get files and related retrieval actions, whose settings and practical limits must be checked in the current connector.

Verify what the action is querying

Confirm the site and document library, then inspect folder restrictions and whether nested items are included. A perfect pagination setting will not return documents outside the selected scope. Also check the connection account's access; the flow cannot enumerate records it is not permitted to see.

Microsoft's Get items and Get files guidance documents the default result behavior and relevant advanced settings. It distinguishes list retrieval from library retrieval, so start by verifying that the action matches the type of source.

Compare a known missing document

Choose an authorized sample you expect to be returned and inspect its folder, metadata and access. If it does not match the filter, increasing the number of retrieved records will not make it appear. Diagnose one concrete omission before broadening the query.

Count the returned array

Inspect the actual output collection rather than the number of rows visible in a collapsed run-history panel. Keep a sanitized count in the test results. Do not copy the entire library response into a public support message to demonstrate an incomplete total.

Review limits as separate controls

Top Count, pagination settings and SharePoint query thresholds solve different problems. A requested count is not a promise that an unselective query can retrieve unlimited records, and enabling pagination is not a reason to ignore the underlying source limits.

The official connector reference provides the current operation parameters and limitations. Use it alongside the retrieval guidance when configuring the installed flow designer. Avoid a copied screenshot that assumes an older action or a different connector.

Observed outputCheck before changing limits
Exactly 100 recordsDefault retrieval behavior and explicit settings
Known subfolder absentFolder scope and nested-item choice
Only one category returnedFilter query and actual source values
Threshold error after increasing countQuery design and supported source limits
Office workers and equipment illustrating a digital document workflow
Illustrative stock photograph of workplace activity, not a product screenshot or a test result. Photo: Pexels.

Test retrieval before enabling work on each file

Keep downstream actions disabled in an authorized test copy. Adjust the applicable settings and compare the returned identifiers with a small, known inventory. Include records that appear late in the expected result, not only the first familiar files.

Use deliberate ordering when order matters

If the business task depends on processing a sequence, specify and validate an appropriate supported ordering. Do not infer a stable business order from whichever results happened to arrive first in one execution.

Account for a changing library

Files can arrive or change while an inventory runs. If completeness is operationally important, define a cutoff or other supported consistency approach with the workflow owner. Pagination alone does not establish a frozen snapshot of a busy document repository.

Add processing with an explicit safety check

Once retrieval is verified, test a small non-destructive downstream action. Record expected and actual processed counts and inspect failures separately. A loop completing successfully for 100 items is not evidence that it handled all eligible documents.

For a library that has grown beyond a broad query's practical scope, refine the retrieval by a meaningful indexed criterion. The SharePoint metadata guide can help define fields that support those business subsets.

Avoid increasing limits indiscriminately on a production flow. Larger outputs can increase execution time and connector activity, and account licensing or request limits still apply. Review those operational constraints before broad deployment.

Have the process owner review the acceptance set independently, including the expected final record. Comparing only against another flow with the same retrieval assumptions can reproduce the omission instead of detecting it.

Keep a short note of the source, filter, folder scope, retrieval settings and acceptance count. Consider the inventory reliable only when representative expected documents appear and the workflow's completeness assumptions are explicit, rather than inferred from an action reporting success.