Use this article when a bulk operation in the scope below did not do what you expected. For example, a mass update or mass delete might show errors, finish without changing records, reject your file, stall, or change records you did not mean to change.
Scope: This article covers operations started from System & Security › Mass Updates, System & Security › Mass Deletes, System & Security › Custom Data Tasks, and Bulk Changes on query and search results.
Start here: Download the results spreadsheet by selecting Click here to download the results. or by using the link in the completion email. Then review the Outcome column, because nearly every answer below starts with that value. If you do not have a results spreadsheet, go to The Run Says It Finished but Nothing, or Only Some Records, Changed.
Warning
A preview can show a Yes/No-style field changing, such as False to True, even when Rhythm does not actually update the record. Do not use the preview alone as proof that the change worked. After you run the operation, confirm the result with the results spreadsheet described in Bulk Operation Previews, Results, and Completion Emails.
Some Records Show Errors
Check: Filter the Outcome column for Error. The reason appears after the colon. Some messages include technical wording or database relationship details, so use the table below to decide what to check and what to do next.
| What you see | Likely cause | Fix |
|---|---|---|
| A reason that names a field or value | Rhythm rejected the new value because it has the wrong type, uses a choice that does not exist, or sends text to a true/false field | Change the value to exactly what the field accepts, then rerun only those rows |
| A reason citing a status or a picklist choice, where the value looks right | Some fields accept a choice only when the capitalization matches exactly, such as a lowercase status token. A membership status update can also depend on other fields you send with it, such as a termination date and a new expiration date | Match the field’s exact text and capitalization. For a membership status, make sure the related date fields are set the way the status expects |
<Record> not found, or "Unable to locate a record with this ID in the database" | The ID does not belong to the record type you chose in your Rhythm account. It is often an ID from another record type, such as an organization ID in a contact update, or it includes extra spaces. Your sandbox is a separate copy from your live account, so an ID such as a membership type ID can validate in one place and not the other. If the ID is a string of digits, opening the downloaded preview file in Excel and saving it can silently reformat that identifier before you upload it for execution | Use IDs for records in the service you chose, confirm the ID belongs to the live account or sandbox where you are running the operation, and avoid editing a numeric-looking ID column in Excel. Keep that column formatted as text |
| Error: with an expression message on every row | The value expression fails for every record, often because a field name is mistyped | Correct the expression, then preview again |
| unsupported array operator | A list value starts with a character other than +, ~, -, or ' | See Mass Update Field Value Rules |
| A required multi-value field reads empty after a remove operation, or a function error appears on another list-type field | Malformed or unsupported multi-select syntax does not always show unsupported array operator. It can also appear as one of these less direct messages | Check the field’s value against Mass Update Field Value Rules, then rerun only those rows |
| "Cannot assign to read only property ... of object '[object String]'" | Updating a multi-select picklist or another array-valued custom field from a file can fail this way instead of applying the update | Rerun the field by itself to confirm the issue, then contact Support if it continues. A direct update to the record through Rhythm’s API can succeed even when the file-based update does not |
| IllegalOperation: Duplicate IDs were found in the request. This is not allowed. in a custom data task | The same record ID appears more than once within one batch of up to 100 rows | Re-import the file in smaller segments so you can isolate and remove the duplicate |
| A mapped column passes the preview but the field is rejected with "is not allowed" only when you process the data | Some custom fields can be selectable and pass mapping and preview, then still be refused at this later stage | Contact Support with the task, the field, and the results if a mapped column is rejected only at this step |
| ObjectInvalid: Invalid owner type undefined on a membership-related update | The message names the owner type, but the actual problem can be a required membership field missing from the update rather than the owner type itself | Check that every required membership field is included and has a value, not only the field you meant to change |
| An "invalid JSON path" error during a custom data task | This message has been observed with a defect affecting how custom fields are handled during import, not only with a bad file or mapping | Contact Support with the file and the exact message if changing the mapping and re-checking the field values does not resolve it |
| "ServerError: Cannot read properties of undefined (reading 'utcOffset')" during an import | The file contains an unsupported or misread date or time value, rather than a problem with the file’s overall layout | Check the format of any date or date-time columns before you re-upload |
| A generic "Unknown Error", sometimes with a 403 shown in browser developer tools, during a custom data task import | The import payload’s JSON can be malformed, for example because editing the file changed its structure | Check and correct the file’s formatting, then retry |
| Numeric-looking values on Certification or Certification Application records are rejected, or saved as dates in a text field | File-based updates to these record types have been reported to misread numeric-looking spreadsheet cells, depending on how the file parses them | Format the column as text in your spreadsheet before you upload, and check the preview’s New column before committing |
| A block of up to one hundred rows with the same error | Rhythm rejected the whole batch, and Rhythm does not retry batches | Rerun only those rows |
| A delete is refused | Rhythm will not delete the record, for example because other records depend on it | Resolve the dependency, then delete again |
- A date is off by one day. Rhythm should keep the date value you provide, so a bulk update or import should not shift it. If a date looks wrong after the change, first check that the spreadsheet cell is formatted as a date rather than text. Also confirm the value does not include a time zone offset that moves midnight to the previous or next day before you treat it as a Rhythm problem.
- You need to remove one value without disturbing the others, such as one characteristic or interest on a set of organization records. Use the list syntax’s remove operator, such as
-[VALUE], instead of a plain value, because a plain value replaces the whole list. See Mass Update Field Value Rules. - Do not work around a rejected postal code by wrapping it in quotation marks. Rhythm saves the quotation marks as part of the value instead of fixing the format. Format the column as text in your spreadsheet so leading zeros and plain digits are kept, then enter the value without quotes.
- A custom field’s option was deleted before it was removed from existing records. The value can remain on those records after the option no longer exists as an active choice. At that point, you may not be able to filter for it to select the affected records for a query-based mass update, especially when the same multi-select field has several selected values. If your own query and mass-update attempts cannot reach every affected record, contact Support; some remaining values need to be cleared on Rhythm’s side.
- Selected records include duplicate rows. A query or search result can list the same record more than once, such as when portal login fields are included in a contact query. Running Bulk Changes over a selection with duplicate rows for the same record can fail with an unexpected error, even though the same change succeeds when you select only one row for that record. If you can, remove the fields that caused the duplicates and rerun the query. You can also export the results with the record’s ID as an output field, remove duplicate rows by ID, and run the mass update from System & Security › Mass Updates against the deduplicated file instead of running it directly from the query results.
- A delete is refused because of a database relationship, and the message does not say what to do next. Open the record and review its Relationships area to see which linked records are involved. Common examples include a document folder that is empty of visible files but still linked to a filing cabinet, a filing cabinet with an entitlement relationship to members through a section, and a record linked through an integration, such as an event linked to a Zoom account. Resolve or remove that relationship, then delete again. Emptying a folder of its visible files is necessary, but it is not always enough if another relationship still points to it. Some record types, such as Documents entitlements and folders, do not support bulk data operations at all; contact Support if normal cleanup still cannot fully remove a blocked or orphaned record.
- An extra header row in an import file. A custom data task import expects one header row. A second header or title row above it can prevent Rhythm from recognizing the columns. Remove any extra rows above the column headers and re-upload.
The Run Says It Finished but Nothing, or Only Some Records, Changed
Work through these in order.
Rows read "Ready For Execution". Those records had nothing to change, so Rhythm skipped them and does not count them. This happens when the new value equals the current value, or when both values are empty. You cannot use a mass update to "resave" records that have not changed, including with a temporary field created just to force a resave. If a record needs to be resaved for a reason other than the field you are updating, such as a Custom Fields section that will not load, open and save that record individually. Some records with fields that populate automatically, such as latitude and longitude generated from an address, may show blank values only because they were imported before that record was ever individually saved. A mass update cannot force the resave that would fill them in. If many records need this kind of refresh and editing them one at a time is not practical, contact Support about a script-based fix rather than trying to force it through a mass update.
Rows read "No Change". Rhythm applied the change, but the saved values already matched what was there.
You changed a field on a different record type. The same field name can exist on several record types, such as a custom field on the membership application and one on the contact. A custom field can also appear on the same record’s screen while belonging to a different application area, such as CRM instead of Membership. Because of that, the mass update must run against the application and service that actually houses the field. The update succeeds on the record type you chose, so the record type you check afterward has not changed. In the preview, an empty Old column for records you know have a value is the warning sign. Rerun against the record type and application that holds the field.
A column was matched to a display field. When a column maps to a read-only display field, the row reports success, but the value does not change. For roles and similar relationships, map to the field ending in
_idsand supply IDs. See Map File Columns to Rhythm Fields. This same file-based approach, mapping to the relationship field and supplying the target record’s ID, is also a way to set a contact’s organization when the normal organization search cannot find the organization you want. Separately, a contact imported with organization text in a display field can show that organization’s name without being linked to an actual organization record. That is different from a real relationship and will not behave like one in queries or mass updates.The run stopped early. If a run reaches its time limit or encounters an unexpected error, it ends without a results spreadsheet and without an email. Depending on the error, you either see Unable to execute mass update or reach the completion step with missing or partial counts, sometimes zero, including a result of zero successes and zero failures with no other explanation. Records written before the stop stay changed. A large custom data task import, such as one with many rows and columns run against activity records, can hit the time limit and still create some records even though their rows show an error, so check whether records were created before you retry. Check a sample of records, or their audit history, to see how far the run got, then split the remaining records into smaller runs. If a large mass update to populate one field across many thousands of records times out, retrying the same change against a smaller batch of records is a good first step to confirm whether the job’s size caused the timeout before you look for another cause. Splitting a large or mixed import, such as separating records that have an email address from those that do not, can also make it easier to isolate and reimport only the records that failed. If a custom data task completes normally but many rows show no change even though the source values clearly differ from the current ones, this can be a product issue rather than bad data; contact Support with the file and results. A row can also fail because the import template is missing a required mapped field, such as an organization ID, or because the record was already in a state the file did not account for.
The records or query results have not caught up yet. Query results can lag behind a record that was just changed, because a query reads from a data source that can take anywhere from a moment to a few minutes to catch up. This lag is more likely right after a bulk change touches many records at once. The same short delay can affect a status change or a banner on a record clearing after a transaction completes, so give it a few minutes before assuming it is stuck. Before assuming an update did not take effect, wait a few minutes and rerun the query or check the record’s audit history, which reflects a save immediately. If a query or report still looks stale well beyond a few minutes, contact Support. In some cases, Support can refresh the affected query data without asking you to resave every record by hand. Removing a banner directly, rather than waiting for it to clear on its own, is something only Support can do. Cleaning up search results after a bulk change touches a large number of records also depends on a background process; if that process has not finished, deleted or changed records can keep showing up in search for a while afterward.
A contact’s organization did not clear. Removing a contact’s organization in a mass update or custom data task can appear to succeed but leave the organization in place when a membership is tied to that contact-organization relationship. Change or remove the membership first if the organization still shows after the update.
A custom field shows the new value on the record but stays blank in query results. This can happen when the saved value’s format does not match what the query expects to display, even though the update itself worked. Check the field’s format, such as a picklist or multi-select value saved as plain text, before assuming the update failed.
Expected relationships, such as group memberships, are missing afterward. If you can put together the correct list of who should be included, you can use a custom data task to reassign the memberships in bulk instead of adding them one at a time. If a combined custom data task that creates contacts and their memberships in one pass fails partway through, split it into two tasks: import the contacts first, then run a second task to import the memberships linked to those contact records.
You changed a field on a different record type. The same field name can exist on several record types, such as a custom field on the membership application and one on the contact. The update succeeds on the record type you chose, so the record type you check afterward has not changed. In the preview, an empty Old column for records you know have a value is the warning sign. Rerun against the record type that holds the field.
A column was matched to a display field. When a column maps to a read-only display field, the row reports success, but the value does not change. For roles and similar relationships, map to the field ending in
_idsand supply IDs. See Map File Columns to Rhythm Fields.A picklist custom field did not take the new value. In some cases, a picklist custom field is not updated even though the row reports success instead of an error. If a picklist column still shows the old value after a run that reported success, check the record directly and contact Support if the value still has not changed.
The run stopped early. If a run reaches its time limit or encounters an unexpected error, it ends without a results spreadsheet and without an email. Depending on the error, you either see Unable to execute mass update or reach the completion step with missing or partial counts, sometimes zero. Records written before the stop stay changed. Check a sample of records, or their audit history, to see how far the run got, then split the remaining records into smaller runs.
Tip
The counts on the completion screen, the email, and the workbook are calculated slightly differently, so they can disagree. Use the per-record rows in the results spreadsheet as the reliable answer.
Your File Was Refused
The preview shows Unable to preview import with a message beneath it.
| Message | Cause | Fix |
|---|---|---|
| "…row(s) with duplicate identifiers…" | The same ID appears more than once. Two blank ID cells also count as duplicates, including blank rows left at the bottom of the spreadsheet after the last record | Sort by the ID column, then remove repeated IDs and any blank rows, especially trailing blank rows at the bottom of the file |
| "…records found where the column column had no value…" | ID cells are blank | Fill in those IDs or remove those rows |
| "Invalid file extension, unable to determine file type" | The file is not .csv, .xls, or .xlsx | Save the file in one of those formats |
| "Unable to read file" | Rhythm could not open the file, often because the link requires sign-in | Upload the file instead of using Use URL |
| "The specified file is empty" / "The input file is empty" | The file has no data rows | Check the file and the first sheet |
| The file looks correct but is rejected anyway, sometimes after rows were deleted while editing | Deleted rows and other workbook formatting artifacts can remain in a spreadsheet even after the visible data looks clean | Copy the visible data into a brand-new spreadsheet, save it, and upload that copy instead |
| The preview fails generically even though the IDs look valid | The file mixes record types, such as an organization row included in a contact update, or it has duplicate IDs that the duplicate-identifier check did not isolate on its own | Keep one record type per file, and remove duplicate rows by ID before you upload |
| "…does not support file-based mass update field mapping." | The service you chose is a lookup type | Choose the record type that holds the field |
| "…does not yet support bulk data operations…" | That record type cannot be changed in bulk. Some record types, such as Documents entitlements and folders, do not support bulk data operations at all. A setting change to Documents entitlements typically applies only to future members, not to existing organizations, which must be added manually | Contact Support if you need it |
Fewer Records in the Preview Than Expected
- Expression selection: A typo in a field name excludes every record without showing an error. A selection expression cannot read custom field values by the custom field’s name, so an expression that tests a custom field finds no value there.
- File selection: Rhythm reads only the first sheet and stops at the first row with an empty column A.
- Not found rows: Use the not-found guidance in the errors table above.
The Screen Is Stuck or Shows an Error
| What you see | What to do |
|---|---|
| Retrieving field information... just a moment... never finishes after the message Unable to load fields for service | At the top of the wizard, select the step 1 symbol, choose the service again, and continue |
| Generating a preview... just a moment... for a long time | Large selections take time, so leave it running. If it never finishes, narrow the selection |
| Unable to execute mass update or Unable to execute mass delete | The commit failed, but Rhythm kept your choices. Try again, or go back and adjust them. If the same change keeps failing from one entry point, such as the Mass Updates wizard under System & Security, try the other bulk-update path, such as Bulk Changes from a query, or the reverse. Some failures are specific to the tool you started from rather than the field values, and the same change can succeed through the other path. Test with one or two records first |
| A Bulk Changes menu item is greyed out with Access Denied | You do not have the permission needed for this operation on this record type. Ask an administrator |
| In a custom data task, the record-type or service selection area is missing, unusable, or does not respond | This has been observed as a permissions issue rather than a browser problem. An Admin or superuser role does not by itself guarantee access to this part of the workflow, and refreshing, clearing your cache, or using a private browser window may not fix it. Ask an administrator to check your profile’s permissions for custom data tasks, and contact Support if access looks correct and the problem continues |
| No records selected | Select at least one record before choosing an operation |
Records You Did Not Expect Were Changed
- The selection is recalculated at commit. With All Records or an expression, Rhythm includes records that match at commit time, even if they were not in the preview.
- The displayed record count was stale. If the selected-record count beside your query or search results looks wrong just before you start a mass update, select the refresh icon next to the count, or rerun the query, to resync the selection before you commit.
- The same operation ran twice. If you commit again, such as after choosing Back from the completion step, Rhythm re-applies every change. With expressions that build on the current value, the effect compounds.
- Two operations overlapped. Each record keeps whichever change reached it last.
- The file’s rows were misaligned. Each row’s values are applied to the record whose ID is on that row. If IDs and values shifted or mixed during editing, the values land on the wrong records. Check that the ID column holds IDs for the record type you chose and that each row belongs together.
To see exactly what happened to a record, look it up in the audit history. There is no undo. If you kept the preview, use the preview’s Old columns to restore values with a new mass update.
When to Contact Support
Contact Support if the causes above do not explain what happened. Include:
- the results spreadsheet, and the preview if you have it;
- the application, service, and selection method you used;
- the file you uploaded, or the IDs or expression;
- the date and time you committed, and the counts shown.