Conversion factors inserted as double are read and converted

tminelgin · Post by **tminelgin** » Fri Jul 13, 2012 5:41 am

I have a process that reads a flat file of UOM data to insert into a table for subsequent processing. The column in the table is defined as DECFLOAT and I insert it as DOUBLE with no issues. Querying the rows in my SQL tool shows all values the same however when I read the table with Datastage some of the values are converted to scientific format, ie exponential. Is there a setting or cast function that will leave these values as is?

ArndW · Post by **ArndW** » Fri Jul 13, 2012 5:43 am

What datatype are you using in DataStage?

(This is a trick question, since "float" or "dfloat" use a binary representation and only convert to display format when you display the value - which would be the point in time when a conversion to an exponential format would occur).

tminelgin · Post by **tminelgin** » Fri Jul 13, 2012 6:15 am

Using DOUBLE as type. It is not only the viewer, when I write these values to a file they are converted.

chulett · Post by **chulett** » Fri Jul 13, 2012 7:18 am

First thing I'd suggest is an exact search here for "scientific", the topic of inadvertent scientific notation has come up several times and I got 40 matches when I just did the same.

ArndW · Post by **ArndW** » Fri Jul 13, 2012 7:23 am

The file is also a "viewer", since the data is being written a text, so a conversion has taken place. You need to specify a column output format to override the default. I like "%f" (see the documentation for sprintf() for an explanation)