Win: 

In order to find your real intention I'll rather ask:

From your post, I'd assume:

- You need a linear regression of an arbitary dataset
  which is in the form of [float(x),float(y)] x 100

- You have 10,000 of these datasets, which you'll pull subsequently
  from a database

- for each of the 100-coordinate data set (10,000), you'll compute some fit
  (a line slope + an intercept) and its corresponding R² to this
  100-coordinate dataset?

- You'll store each of the (10,000) results, together with Record-ID, slope,
  intercept and R² subsequently into a file (10,000 rows), which is
  your new table?

Just asking to be sure ... ;-)

Regards

mwa