langfuse/langfuse · error · LangfuseNotFoundError

The dataset item with id ${input.id} already exists in a dat

Error message

The dataset item with id ${input.id} already exists in a dataset other than ${datasetIdentifierForLog}

What it means

On upsert of a dataset item, the (id, projectId) uniqueness path fires when the supplied id already exists but is attached to a different dataset than the target. The service logs a warning and returns 404-style NotFoundError rather than moving the item between datasets.

Source

Thrown at web/src/features/datasets/server/publicDatasetService.ts:679

      ...transformDbDatasetItemDomainToAPIDatasetItem({
        ...datasetItem,
        datasetName: datasetItem.datasetName,
        status: datasetItem.status ?? "ACTIVE",
      }),
      mediaReferences:
        mediaReferences.get(datasetItemMediaReferenceKey(datasetItem)) ?? [],
    };
  } catch (error) {
    if (error instanceof Prisma.PrismaClientKnownRequestError) {
      if (error.code === "P2025") {
        // this case happens when a dataset item was created for a different dataset.
        // In the database, the uniqueness constraint is on (id, projectId) only.
        // When this constraint is violated, the database will upsert based on (id, projectId, datasetId).
        // If this record does not exist, the database will throw an error.
        logger.warn(
          `Failed to upsert dataset item. Dataset item ${input.id} already exists for a different dataset than ${datasetIdentifierForLog}`,
        );
        throw new LangfuseNotFoundError(
          `The dataset item with id ${input.id} already exists in a dataset other than ${datasetIdentifierForLog}`,
        );
      }

      if (error.code === "P2002") {
        // Unique constraint violation on (id, projectId, validFrom).
        // This can happen when concurrent requests try to update the same dataset item
        // and create versions with the same timestamp.
        logger.warn(
          `Failed to upsert dataset item due to version conflict. Dataset item ${input.id} was modified concurrently.`,
        );
        throw new LangfuseConflictError(
          `Dataset item ${input.id ?? "new"} was modified concurrently. Please retry the request.`,
        );
      }
    }

    throw error;

View on GitHub (pinned to 59d92c7cf3)

Solutions

1. Use a dataset-scoped id scheme (e.g. `${datasetName}:${itemId}`) or omit id to let Langfuse generate one
2. Delete the item from the old dataset first if you intend to move it
3. Check which dataset owns the id: GET /api/public/dataset-items/:id

Example fix

// before
await datasetItems.create({ id: "item-1", datasetName: "dataset-b", ... });
// after
await datasetItems.create({ id: "dataset-b:item-1", datasetName: "dataset-b", ... });

const existing = await client.fetch(`/api/public/dataset-items/${id}`);
if (existing && existing.datasetName !== targetDataset) id = `${targetDataset}:${id}`;

catch (e) { if (is404WithMessage(e, /already exists in a dataset other than/)) useNamespacedId(); else throw e; }

When it happens

Trigger: POST/PUT /api/public/dataset-items with an explicit id that already exists under another dataset in the same project (e.g. re-using an id from dataset A when writing to dataset B).

Common situations: Copy/sync scripts that reuse deterministic item ids across datasets; renaming a dataset and recreating it under a new name while re-importing items with the same ids.

Related errors

• Dataset item not found
• A queue with this name already exists.
• Dataset id must not be empty
• Dataset name not valid. ${validation.error.message}
• ${duplicateErrorMessage}

AI-assisted analysis of langfuse/langfuse@59d92c7cf3 (2026-08-27). Data as JSON: /api/errors/e51936fdf56c8b8c. Report an issue: GitHub.