Please start any new threads on our new
site at https://forums.sqlteam.com. We've got lots of great SQL Server
experts to answer whatever question you can come up with.
| Author |
Topic |
|
malakai
Starting Member
8 Posts |
Posted - 2004-05-28 : 12:43:27
|
| I have to load a large amount of data from CSV's which the source documents many columns as being FLOAT(53) and which represent currency. My current process brings these over as FLOAT, but i'm struggling with performance, indexes, and row sizes. Some of the entities i'm transforming and loading have over 255 columns. 50% of which are FLOAT. I'd love to remove as much non-deterministic columns as possible.What should I consider when converting from FLOAT(53) to Money? Money obvisouly has a different scale then FLOAT, how does SQL handle conversion by default? Is it better to truncate or round? I'm ignoring right now penny issues, because this system is simply reporting data and analyzing at a much higher level. Is it worth the hassle factor out the FLOATs? If I went Numeric with specific precision and scale do I buy myself performance or better responding indexes vs Money type?I have a gut reaction to remove as many of the FLOAT columns as I can, but i'm not sure I can qualify why I feel this. I would appreciate any comments or experiences.-frank |
|
|
Page47
Master Smack Fu Yak Hacker
2878 Posts |
Posted - 2004-05-28 : 13:34:41
|
| When you convert a float to a money, SQL will round to 4 decimal places.Jay White{0} |
 |
|
|
derrickleggett
Pointy Haired Yak DBA
4184 Posts |
Posted - 2004-05-28 : 13:37:25
|
| And you need to do so before doing your analysis. I hate to even think about what using floats would do to your analysis. FLOAT is not acceptable to use for currency. Whatever "developer" designed the other system needs to change it.MeanOldDBAderrickleggett@hotmail.comWhen life gives you a lemon, fire the DBA. |
 |
|
|
|
|
|