# Striim 3.9.6 documentation

#### MSSQLReader data type support and conversion reference

MSSQL type

TQL type

notes

bigint

long

binary

byte[]

not included in before array for UPDATE or data array for DELETE operations: see cautionary note below

bit

string

char

string

date

string

datetime

string

datetime2

string

datetimeoffset

string

decimal

string

float

double

image

byte[]

not included in before array for UPDATE or data array for DELETE operations: see cautionary note below

int

int

money

string

nchar

string

ntext

string

not included in before array for UPDATE or data array for DELETE operations: see cautionary note below

numeric

string

nvarchar

string

nvarchar(max)

string

included in before array for UPDATE operations only if value is changed by the update

real

float

smalldatetime

string

smallint

short

smallmoney

string

text

string

not included in before array for UPDATE or data array for DELETE operations: see cautionary note below

time

string

timestamp

byte[]

tinyint

short

udt

string

uniqueidentifier

string

varbinary

byte[]

not included in before array for UPDATE or data array for DELETE operations: see cautionary note below

varbinary(max)

byte[]

not included in before array for UPDATE or data array for DELETE operations: see cautionary note below

varchar

string

varchar(max)

string

included in before array for UPDATE operations only if value is changed by the update

xml

string

included in before array for UPDATE operations only if value is changed by the update

### Caution

When all tables being read have primary keys and none of those primary key columns is of type binary, image, ntext, text, varbinary, and varbinary(max), you will not encounter the following issue.

When replicating MSSQLReader output using DatabaseWriter, if one or more of a table's primary key columns is of type binary, image, ntext, text, varbinary, or varbinary(max), or if a table has no primary key and one more columns of those types, UPDATE or DELETE operations may erroneously be replicated to more than one row.This may result in additional errors when subsequent operations try to update or delete the missing or incorrectly updated rows.