[Bug]: File Metadata does not scan for decimals properly

### Apache Hop version?

2.12 +

### Java version?

17.0.16

### Operating system

Linux

### What happened?

I have a file that has currency values - (example: 12423.25) File Metadata is reading it as a Number with length 15 and precision = 1 and mask #.# - which subsequently drops off any number past the first decimal.

Also - if there is an embedded quote (when all fields are enclosed in quotes), the transform may fail if it is a large dataset. It would be nice to handle that situation - **which this also breaks the enclosure from being identified even if the transform does not throw an error**

File attached

### Issue Priority

Priority: 3

### Issue Component

Component: Transforms

Start with the File Metadata transform using the attached file in the stated Hop 2.12 and Java 17 environment. Reproduce decimal inference and embedded-quote behavior on a large dataset; done means decimal length, precision, and mask retain values such as 12423.25, while quoted fields remain correctly enclosed without transform failure.

Written by the indexing model from the issue text.