Implications of Large Datasets (more than 100M rows) in Wave Analytics
|Knowledge Article Number||000240394|
A single dataset should not exceed 100 million rows of data.
API queries have a 120 second timeout. Wave has been optimized to return queries on datasets of up to 100M records within that window. Datasets that exceed 100M rows may cause related queries to exceed the timeout threshold.
From the API Developer Guide: "Queries that take longer than two minutes to process will be timed out."
From Wave Analytics Limits: "For best performance, we recommend datasets to be in the range of 100M records. Several items, such as the size of the records, and cardinality of measures, can significantly or adversely affect performance."
Impact of Oversize Datasets
Split large datasets into manageable chunks and refactor dataflows and dashboards to accommodate.