Is there any way to skip rows when I retrieve from Azure table storage?

It is still not possible to skip rows. The only navigation construct supported is top. The Table Service REST API is the definitive way to access Wndows Azure Storage, so its documentation is the go-to location for what is or is not possible.

What you're asking here is possible using continuation tokens. Scott Densmore blogged about this a while ago to explain how you can use continuation tokens for paging when you're displaying a table (like what you're asking here with DataTables): Paging with Windows Azure Table Storage. The blog post shows how to show pages of 3 items while using continuation tokens to move forward and back between pages:

Besides that there's also Steve's post that describes the same concept: Paging Over Data in Windows Azure Tables

Yes (kinda) and no. No, in the sense that the Skip operation is not directly supported at the REST head. You could of course do it in memory, but that would defeat the purpose.

However, you can of course actually do this pattern if you structure your data correctly. We do something like this ourselves. We align our partition key to the datetime and use the RowKey as a discriminator. This means we can always pinpoint the partition range we are interested in and then Take() some amount of data. So, for example, we can easily Take() the first 20 rows per hour by specifying a unique query (skipping over data we don't want). The partion key is simply aligned per hour and then we optionally discriminate further using the RowKey - finally, we just take data. When executed in parallel, this works just dandy.

Again, the more technically correct answer is NO. However, you can approximate it cleverly using the PK and RK.