Script as INSERT chooses inappropriate FLOAT datatypes
Mark_G
Posts: 1 New member
in SQL Prompt
A word of warning...
When you take a results grid which contains float values, and run a "Script as Insert" on the data, the generated SQL seems to always choose the data type "float(8)" to represent the inserted data.
The trouble with this is, the 8 used here means 8 mantissa bits (which SQL Server rounds up to a single-precision float, according to the docs), rather than the 8 bytes you might assume - i.e. a double-precision float.
This means, Script as INSERT loses a huge amount of float precision when the SQL is executed.
Cheers,
Mark
When you take a results grid which contains float values, and run a "Script as Insert" on the data, the generated SQL seems to always choose the data type "float(8)" to represent the inserted data.
The trouble with this is, the 8 used here means 8 mantissa bits (which SQL Server rounds up to a single-precision float, according to the docs), rather than the 8 bytes you might assume - i.e. a double-precision float.
This means, Script as INSERT loses a huge amount of float precision when the SQL is executed.
Cheers,
Mark
Tagged:
Comments