It shouldn't be happened. Are you using ZN() to handle the null value in your measure? Is rank measure a calculated field I guess? Can you share a screen shot of your worksheet? If you've used any calculated measure in your chart then share the formula of that measure.
If you actually have the year 2015 I'm surprised the line is joined. If you don't have any entry for 2015 at all - i.e. Tableau is creating it (or displaying it actually as no value will have been created for it) there's a way around this.
Make your Year a discrete (blue) value and 'Show missing values'
Andrew If there is no entry for year 2015 then in discrete/continuous line also it will not join. See this with continuous
If the user has handled nulls using ZN() in the measure then it'll join the join with deep spike at zero line.
Correct me If I'm wrong.
I didn't actually use the ZN at all Mahfooj Khan, however your example makes it clear handling nulls in that way joins the line. As an FYI as much as anything, as I'm not sure what the underlying data of the OP actually looks like, these were the scenarios I ran and the results:
With the date in the data but with a null Position:
When charted Continuous it gives:
WIth the dataset altered, where we have no entry at all for 2015:
When charted Continuous it gives different results to the previous example:
This is why I said what the data actually looks like behind the scenes matters, whether the year does actually exist in the data with a null position or whether it isn't in the underlying data at all.
Yeah that might be the case. You're right without seeing the underlying data we can't predict what may be the case.