Deep Insights in Microsoft .NET

If you need to do this, use HashSet<T> which is 100X faster!

If you want to migrate a lot of data which means running updates to existing rows, it is 100X faster to first change the objects then run a single save changes call, instead of doing row by row call. Also, when batching such operations say by 5000 rows, re-initialize the entities object each time, otherwise it will cache everything progressively getting slower over time.

I just ran a benchmark today. DataContractSerializer runs pretty fast, but The DataContractJsonSerializer can serialize faster. Note however that the DataContractSerializer can deserialize faster than the Json. Deserialization times: Binary: 198 ms DC: 20 ms Json: 74 ms Serialization times: Binary: 191 ms DC: 0.3 ms Json: 0.2 ms When the size of data increases, Binary slows down faster than other techniques. Also, the size of the data is significantly smaller for Json and DC.

This comes up again and again as a thorny and persistent problem. I got this new way of doing this from my friend. SELECT alias1. PKCol1 , alias1.Col2 FROM table1 alias1 WHERE alias1.Col3= 11 AND alias1.Col4= 43 AND       NOT EXISTS (SELECT 'x' FROM table2 alias2 WHERE alias2. PKCol1 = alias1. PKCol1 )