Include blank values in a scan

[wchillman]

Here is the situation. I ran a simple scan that looked for:

plot scan = [Criteria] > 20;

Of course, it worked perfectly by excluding all entries for which the criteria were less than or equal to 20. However, it also excluded entries that were blank (actually, when I ran my cursor over them it indicated “N/A”.

Now, what do I have to do to get these blank entries included with the original scan?

 

[Svanoy]

Based on the code provided:

plot scan = IsNaN([Criteria]) or [Criteria] > 20;

This will include all N/A and greater than 20.

 

[wchillman]

Ruby,

Thank you for your reply. It sure looked like it was going to work! However, I am at a loss here to explain why it doesn't. Let's narrow the playing field to:

plot scan = Close > 2 AND Close < 3 AND Volume < 2000;

Now add this as a column to the scan:

def LB = 252;
def HH = Highest(High,LB);
def LL = Lowest(Low, LB);
plot DR = 100*(HH/LL-1);

You will see many "N/A" entries. If I add the following scan,

def LB = 252;
def HH = Highest(High, LB);
def LL = Lowest(Low, LB);
def Criteria = 100*(HH/LL-1);
plot scan = IsNaN(Criteria) or Criteria > 20;

I lose all the "N/A" entries. Obviously, all those entries are stocks that have less than 252 days of trading history. What do I need to do to get them included?

 

[Joshua]

Having a NaN value appear on either side of a logical (or), comparison (>), or arithmetic operator will cause the entire expression to resolve to NaN. If Criteria's value is NaN, then NaN > 20 resolves to NaN, and then Anything Or NaN will also resolve to NaN. This can only be handled by a conditional expression, and in the correct order. You must ensure that the value isn't NaN (if) before attempting to resolve it (else if).

Plot Scan =
if isNaN(Criteria) then Yes
else if Criteria > 20 then Yes
else No;

This will not work:
Plot Scan =
if Criteria > 20 then Yes
else if isNaN(Criteria) then Yes
else No;

 

[wchillman]

Joshua,

Thank you for your explanation. Here’s my scan with your revision. First, I narrowed the playing field with this scan:

plot Scan = Close > 2 AND Close < 3 AND Volume < 2000;

It yields over 100 stocks and many of them show a result of “N/A”. Next, if I add this revised scan,

def LB = 252;
def HH = Highest(High, LB);
def LL = Lowest(Low, LB);
def Criteria = 100*(HH/LL-1);
plot Scan =
if isNaN(Criteria) then Yes
else if Criteria > 20 then Yes
else No;

I still lose all the “N/A” entries. What am I missing?

 

[Joshua]

There is a hard left edge at the absolute beginning of the existing data, it behaves quite differently than the right side expansion area. Nothing beyond that edge can be retrieved or evaluated at all, attempting to do so doesn't even produce a NaN value.

I would suggest determining when you're within 252 bars of the right side edge, by using an IsNaN() check against a forward offset. Once you're within that range, start tabulating the highest high and lowest low manually.

If there are less than 252 bars, you'll end up with the highest and lowest out of what ever number of bars there are. You can also count the bars to determine if there are less than 252, and handle that how ever you want.