问题
I have a C# RestApi which need to extract information of records from database using ADO by executing queries like this:
declare @id1 int
set @id1 = (select id from myTable where col1= x1, col2 = y1, col3 = z1 from mappingTable)
declare @id2 int
set @id2 = (select id from myTable where col1= x2, col2 = y2, col3 = z2 from mappingTable)
declare @id3 int
set @id3 = (select id from myTable where col1= x3, col2 = y3, col3 = z3 from mappingTable)
.
.
.
declare @idN int
set @idN = (select id from myTable where col1= xN, col2 = yN, col3 = zN from mappingTable)
select @id1,@id2,@id3, ..., @idN
I run above query which runs N queries inside it using ADO.NET SqlCommand
and read the results. I have two questions:
- Does running each of queries using separate
SqlCommand
lead to performance downgrade or not? Usually in I/O tasks, executing many small I/O tasks have lower performance than running all of them in one batch I/O tasks but I have no idea about the same scenario in Databases and ADO. - Are there any better ways to extract the same result with a better SQL query? In other words can I write this query other way to run it with better performance?
Note: In above queries columns and tables are the same in all queries only values in where clause are modified.
回答1:
Use a table valued parameter to store corresponding values of x
, y
and z
.
Use a single query where you inner join your mappingTable
with that table valued parameter.
In SQL:
CREATE TYPE XYZ As Table -- Do not call it XYZ!!!
(
x int,
y int,
z int -- I'm using ints because it's simple - you need the data types of your columns here
)
In c#, create a data table that corresponds to this type:
var dt = new DataTable();
dt.Columns.Add("x", typeof(int));
dt.Columns.Add("y", typeof(int));
dt.Columns.Add("z", typeof(int));
populate that data table with the correct values, and then you send the data table to SQL Server as a parameter of type SqlDbType.Structured
:
cmd.Parameters.Add("@tvp", SqlDbType.Structured).Value = dt;
To use it in a query:
SELECT id
FROM mappingTable
JOIN @tvp
ON col1= x
AND col2 = y
AND col3 = z
This will return a recordset containing a single column with all the ids you need.
Note: Code was written directly here - there might be some typos.
来源:https://stackoverflow.com/questions/59298365/c-sharp-ado-optimize-query-and-its-performance