For usability, the table APIs allow removal of rows and columns by providing lists of potentially unordered nodes. To improve performance, the table schema code has now been updated to group contiguous ranges of these nodes and do batch deletions from their corresponding lists, rather than removing 1 at a time. This reduces the number of local ops. All of the edits are performed in a transaction (when the tree is hydrated), so this will only impact the performance of local edits, and not that of collaborators applying remote ops.