About count (1) and COUNT (*) Efficiency differences and differences

In the previous work in the hospital, for SQL optimization, many people think that count (1) is more than count (*) faster, let me always think it is true, and later found that it is not actually. In the same table query, without the primary key case, the query speed difference is not large, into the background, will find that, in fact, Count (1) in the background is automatically converted to count (*)

Here are the other Daniel's conclusions, quote

There is also a count (column), in which the Query method and the same point of Count (*) are all non-null, so their differences are also here, when a field such as DD is null in one piece of data, and the other is not NULL, COUNT (*) and COUNT (DD) The query results are different, COUNT (*) is 1,count (DD) is 0. Specific use to see the situation