DataFrame.OrderBy(string columnName) does not perform stable sorting!

Stable sorting preserves the order of entries in the input when their sorting keys are same. Please see the below link:
https://stackoverflow.com/questions/1517793/what-is-stability-in-sorting-algorithms-and-why-is-it-important

As an example:
Before sorting...

```
df
A B C
9 18 28
6 11 25
6 10 23
3 16 26
6 13 21
3 19 22
3 11 28
6 17 20
```

After sorting...

```
df.OrderBy("A")
A B C
3 16 26
3 11 28
3 19 22
6 11 25
6 10 23
6 17 20
6 13 21
9 18 28
```

The three rows where column `A` = `3`, should be ordered with column `B` as `16,19,11` (not `16,11,19`).
The four rows where column `A` = `6`, should be ordered with column `B` as `11,10,13,17` (not `11,10,17,13`).

If we get a stable sorting with `OrderBy` function, we would have a work-around for multi-column sorting which is not yet available according to [#5758](https://github.com/dotnet/machinelearning/issues/5758) and [#5649](https://github.com/dotnet/machinelearning/issues/5649). For those interested in the workaround, please see [this question on stackoverflow](https://stackoverflow.com/questions/24573337/c-sharp-deedle-sort-by-multiple-columns).

Thanks!