From the REST-API side of things, this requires a new endpoint to do the result comparison. Doing the diff computation in the browser is too inefficient given the fact that the whole result would have to be turned into JSON and then a diff be generated from that, the amount of browser memory required for this would be very high. Calculating an actual diff over the results might also not be necessary as that would potentially require storing both results in memory to compute it and if a natural "sort key" is available, is not needed. For the simple case of comparing two tables, comparing the results of the table ordered by the primary key should be simple enough to implement so that it can be done automatically. For arbitrary queries, the user would have to tell which field to use as the fake "primary key" that is used to compute the difference of the two results.
The API should reuse the existing code from the ETL feature that allows two connections to be used at the same time. A new /sql/diff endpoint would then be given two connections and an SQL query and the returned result would be the diff of the query results.
Hi @Duong
Can we implement for both data compare and structure compare features would be good. I mean it should compare both data and compare all the schema objects (Schema/Table/Procs/Triggers/Any) from source to destination.