October 2019
Intermediate to advanced
520 pages
13h 5m
English
It's worth noting that designing a Bigtable schema is very different from designing a schema for a relational database. There are no secondary indices as each table only has one index. Rows are sorted lexicographically by row key (the only index), from the lowest to the highest byte string. If you have two rows in a table, it's possible that one row will be updated successfully while the other row will fail. Therefore, we should avoid designing schemas that require atomicity across rows. Poor schema designs can be responsible for overall poor Bigtable performance.
Read now
Unlock full access