Improve Insert into table performance in sql - sql-server

Question

There are a couple of things confusing me at this level.
I have a table with around 40 columns out of which atleast 35 are in where clause at different times in single execution of a procedure.
When these 35 columns are passed a value via stored procedure, the stored procedure call their respective inline TVF's which in turn call a common multiline TVF.
I need to know if I shall consider creating indexes for all these 35 columns (Though I have serious doubts if that can help, but please tell me I am wrong if it does. )
I am inserting data into a Temporary table . This insert goes on for number of parameters passed to stored procedure and execution plan shows it takes quite a considerable amount of time. Is there a way I can improve the performance here ?
The insert query looks like this :
INSERT INTo #Temp2
(RowNumber,ValFromUser,ColumnName,ValFromFunc,FuncWeight,percentage)
SELECT RowNumber,#firstname,'firstname',PercentMatch,
#constVal,PercentMatch * #constVal FROM dbo.MatchFirstName(#firstname)
:
Execution plan is attached :
execution plan
Table with large number of columns is as follows :
create table Patients
(
Rowid int identity(1,1),
firstname nvarchar(20) not null,
middlename nvarchar(20),
lastname nvarchar(20)not null,
DOB Date,
SSN nvarchar(30),
ZIP nvarchar(10),
[State] nvarchar(2),
City nvarchar(20),
StreetName nvarchar(20),
StreetType nvarchar(20),
BuildingNumber int,
Aptnumber nvarchar(10),
patientnickname nvarchar(20),
patientsMaidenlastname nvarchar(20),
fathersFirstName nvarchar(20),
fatherslastname nvarchar(20),
mothersfirstname nvarchar(20),
motherslastname nvarchar(20),
mothersMaidenlastname nvarchar(20),
citizenship nvarchar(20),
nationality nvarchar(20),
ethnicity nvarchar(20),
race nvarchar(20),
religion nvarchar(20),
primarylanguage nvarchar(20),
patientmrn nvarchar(30),
hospitalname nvarchar(30),
Medicaidid nvarchar(10),
pcpnpi nvarchar(10),
phonenumber nvarchar(15),
email nvarchar(30),
CreatedAt datetime default getdate(),
ModifiedAt datetime DEFAULT getdate(),
CreatedBy nvarchar(128) default SUSER_NAME(),
ModifiedBy nvarchar(128) default SUSER_NAME()
);
Temporary table looks like this :
create table #Temp2
(
Rownumber int not null,
ValFromUser nvarchar(30),
ColumnName nvarchar(30),
ValFromFunc decimal(18, 4),
FuncWeight decimal(18, 4),
Percentage decimal(18, 4) not null,
);
ResultsStored table :
create table ResultsStored
(
Sno int identity(1,1),
SearchSerial int,
StringSearched varbinary(8000),
RowId int,
PercentMatch decimal(18,4),
CreatedAt datetime default getdate(),
ModifiedAt datetime default getdate(),
CreatedBy nvarchar(128) default SUSER_Name(),
ModifiedBy nvarchar(128) default SUSER_NAME(),
HashedKey binary(16)
);

Answer 1

Indexes speed up (sometimes) SELECTs. But indexes slow down INSERTs (and also DELETEs, UPDATEs etc.). So it's a bad idea to have too many indexes.
Quite often SELECT is able to use only one index (or even zero). So all other 34 indexes are nothing of help. Keep only those ones your SELECTs really use.
As you say, you have about 40 columns in a table with at least 35 columns are being mentioned in distinct 'WHERE'-clauses. That is your table not just big, but, which is far worse, it has too many potential keys. It's a very bad design. You need to split it to several tables. Read about normalization.