No. In Excel for Microsoft 365, a normal Power Query refresh reruns the query against its source and updates the full result; it does not automatically identify and append only rows that arrived since the last refresh. You can make refresh more efficient or build a deliberate incremental process, but Excel does not provide an automatic incremental-refresh policy for ordinary Power Query tables.
What a Power Query refresh actually does
A Power Query is a set of instructions for connecting to data and transforming it. Those instructions might import a workbook table, combine files in a folder, filter records, change data types, and select columns. When you refresh, Excel runs those instructions again and loads the resulting data to the destination you chose, such as a worksheet table or the Data Model.
That distinction matters: the query definition is saved, but the query result is not a permanent log of just the new records. If a source table contained 10,000 rows on Monday and 10,200 on Tuesday, a regular refresh produces the result of the query over the Tuesday source. If the query selects all rows, that will normally include all 10,200 rows, not just the latest 200.
Refresh is also not the same as manually adding rows to the bottom of the query output. The output is managed by Power Query. On refresh, Excel updates the query table to match the new result, so hand-entered changes in that output area can be lost, displaced, or prevented from being entered. Keep manually maintained notes or adjustments in a separate table and combine them deliberately if needed.
Why a normal refresh reads the full result
Power Query generally evaluates the steps in its query from the source each time it refreshes. The query does not automatically remember the last row it loaded, compare every record with a stored copy, or maintain a database-style change log. If the query asks for all qualifying records, the refreshed result is all qualifying records.
Exactly how much data travels over the connection depends on the source and the query. For example, a relational database may execute a filter on its own server and send Excel only the matching rows. Power Query may be able to fold supported transformation steps into a source query, a process called query folding. A folded filter can reduce transfer and work, but it is still not the same as Excel retaining the previous result and requesting only newly added records.
Other sources behave differently. Reading a local CSV or a collection of files may involve opening and processing those files again. A database or web service may have its own caching, indexing, or API behavior. Excel does not guarantee that a refresh is incremental merely because the source connection is fast or the result looks similar to the previous one.
Can Excel Power Query refresh only new rows?
Not automatically for a standard query loaded to a worksheet. Excel does not offer the same built-in incremental-refresh policy commonly associated with Power BI semantic models, where historical data can be partitioned and recent periods refreshed under a defined policy. In Excel, you need to arrange the source or workflow so that it supplies only the records you intend to process, or use a system designed to manage incremental loading.
A common approach is to filter at the source using a date, sequence number, or other reliable watermark. For instance, a database query might return transactions from the most recent date onward. But a filter such as “date is after the last refresh date” is only safe if the process stores that date reliably and accounts for late-arriving or corrected records. A newly imported transaction may have an older transaction date, so a strict date cutoff could miss it.
Another approach is to query a source that already contains only the new batch, then append that batch to a durable destination using a separate controlled process. A normal Power Query worksheet refresh does not itself preserve the old output and append a new batch to it. If the query’s source is a folder containing every historical file, refreshing the folder query normally rebuilds the combined result from the files it processes.
Practical ways to reduce refresh time
If the goal is a faster refresh rather than a new-rows-only result, start by reducing unnecessary work. Apply filters and remove unused columns as early as practical, especially when the source supports query folding. Check whether the source query can do the filtering itself; for databases, a well-designed view or query can limit the records and fields Excel needs to retrieve.
- Import fewer columns. Drop fields that are not used in the final report. Wide records can consume time and memory even when the row count is moderate.
- Filter early. Restrict dates, regions, statuses, or other relevant fields near the start of the query, where the connector can potentially translate the step to the source.
- Use a suitable source. A database table or curated view may be more efficient than repeatedly combining many large flat files.
- Keep transformations purposeful. Avoid expensive row-by-row operations or repeated sorts when the output does not need them.
- Test the result. Confirm that filters do not exclude valid historical corrections, late-arriving data, or records with blank dates.
Whether a particular step folds depends on the connector, source, and transformation. Power Query provides indicators for query folding in some authoring experiences, but not every connector exposes the same diagnostics. If a step stops folding, later operations may be performed by the local Power Query engine after a larger set of source data has been retrieved.
Designing a dependable incremental workflow
For an actual incremental process, decide first what “new” means. It could mean a newly created record, a changed record, or a new file. These are different requirements. A timestamp can identify recent changes only if the source updates it consistently; a unique increasing ID can identify later inserts but may not reveal edits to older records.
A reliable design usually needs a persistent destination outside the refreshed query output, a way to identify records already loaded, and rules for updates and deletions. A database with a staging table and a unique key is often better suited to this job than a workbook table. An ETL tool or a Power BI model with an appropriate incremental-refresh setup may also be a better fit for large or regularly updated history.
For a small, controlled Excel process, one option is to have the source system export dated batches and use a separate append process to store them. Keep a stable record identifier, check for duplicates, and decide how to handle corrections to past records. A rolling overlap window can help catch late changes—for example, reloading a recent period rather than only records after a single cutoff—but it requires a key-based deduplication or update step. Power Query can help transform and combine data, but the persistence and append logic must be designed rather than assumed.
Refreshing a worksheet query and managing its settings
In Excel for Microsoft 365 on Windows, use Data > Refresh All to refresh workbook connections, or open Data > Queries & Connections, right-click a query, and choose Refresh to refresh that query. A query loaded as a worksheet table updates that table to reflect the query result; it is not an append-only archive. If a refresh changes the number of rows, formulas or references around the table may need to account for the changed table size.
To review refresh behavior, open the Queries & Connections pane, right-click the query, and choose Properties. Depending on the connection, available settings can include refreshing when the workbook opens or enabling background refresh. These settings change when or how a query refreshes, not whether Excel automatically refreshes only new rows. Refreshing in the background can let you continue working, but results may not be available immediately while the operation is running.
Excel for Mac and Excel for the web have different Power Query and connection capabilities, and the available connectors and refresh controls can vary by platform and source. If a workbook is shared across platforms, verify that the intended users can refresh its particular connection. A refresh setting is not a substitute for confirming that credentials, network access, and source permissions are available on the computer doing the refresh.
Common misconceptions and refresh problems
“The query only added 200 rows, so it must have read just 200.” The output difference does not show how much data was read. Excel may have evaluated a larger source and produced only 200 additional qualifying rows. Query folding can reduce the work, but the row count added to the table is not proof of incremental loading.
“Refresh All will append the new rows.” Refresh All runs the workbook’s refreshable connections. It does not turn query results into a transaction log. If an append-only history is required, the workflow needs a separate persistent destination and append or merge logic.
“The refresh failed, so the query must have changed.” A refresh can fail because a file moved, a column was renamed, credentials expired, a server is unavailable, or the source schema changed. Check the error details and the source before changing transformation steps. A query that expects a column named Amount, for example, cannot complete if the source now labels it Sales Amount.
Large refreshes can also appear stalled, particularly when many files are combined or a source is slow. Check whether Excel is still processing, whether background refresh is enabled, and whether other queries depend on the one being refreshed. Do not edit or delete the query output as a workaround until you understand how the query is loaded; the next refresh can recreate or replace that output.
Choosing the right approach for your data
For a modest table where the source is readily available and a complete current result is useful, an ordinary refresh is usually simplest and least error-prone. It gives a consistent result based on the current source and avoids maintaining a separate incremental state. For a report that filters a large database, push appropriate filtering to the database and keep the query result limited to what the workbook needs.
For a very large history that grows continuously, ask whether Excel is the right place to store and process the history. A database, a managed data pipeline, or a Power BI model may handle partitioning, change tracking, and long-term storage more appropriately. The right choice depends on source size, refresh frequency, correction patterns, governance, and who must maintain the process—not simply on whether a workbook can technically load the rows.
Before changing a workflow, document the expected behavior: whether old records can be edited, whether deletions must appear, how duplicate IDs are handled, and how far back late data can arrive. Those requirements determine whether a simple full refresh is safest or whether a carefully managed incremental design is justified.
Frequently Asked Questions
Does Power Query overwrite the old Excel table when I refresh?
It updates the query output to match the refreshed result. Treat the output as query-managed data, not as a place to preserve manual edits or build an append-only history.
Can I make Excel Power Query append new rows automatically?
Not as the default behavior of a worksheet-loaded query. You need a separate design that persists prior rows and appends or merges a new batch, or a source that already provides the desired incremental result.
Does query folding mean Power Query refreshes only new data?
No. Query folding can let a source perform filters and transformations efficiently, reducing data transferred to Excel. It does not by itself preserve old output or track which rows are new since the previous refresh.
Will a date filter make the query incremental?
It can limit the records returned, but a fixed or manually chosen cutoff can miss late-arriving records and changes to older records. A dependable process needs a maintained watermark and a plan for corrections, duplicates, and deletions.
Why does Refresh All take a long time?
It may be reading many rows or files, performing transformations locally, waiting for a slow source, or refreshing dependent queries. Reduce unnecessary columns and rows, check source availability, and see whether supported filters can be applied at the source.
Related
- Excel Keyboard Shortcuts: Essential Keys and Tips Guide
- Excel MAX Between Dates: Find the Highest Value in a Range
- Excel MAX Between Two Dates: Find the Highest Value
Try it: the Date Difference Calculator: days, months, years and weekdays between two dates, with the DATEDIF formulas.
More in How-to basics
- How to Create a Postcard Template in Microsoft WordCreate a postcard template in Word by setting a custom page size, arranging the front and back on…
- Excel Page Break View Without Page Numbers: What to DoIn Excel for Microsoft 365 on Windows, you cannot hide the large “Page 1,” “Page 2,” and similar…
- Track a Budget in Google Sheets: A Practical GuideTo track a budget in Google Sheets, make one tab for transactions and another for your monthly plan,…
- How to Hide Columns in Excel Without Deleting DataTo hide columns in Excel for Microsoft 365 on Windows, select the column letters, right-click the selection, and…
- Excel 3-Digit Number Format: Show Leading ZerosTo display numbers as three digits in Excel, apply the custom number format 000: select the cells, press…
- How to Add a Drop-Down List in Excel: Step-by-StepTo add a drop-down list in Excel for Microsoft 365 on Windows, put the choices in cells, select…
Sources: Function behaviour and menu paths: Microsoft Support documentation for Excel. Method & sources
Spotted an error in this guide? Let us know.