Hi Art, do I initialize the array with missing values.. ?  using retain ? What I'm trying to do is.. run the data step through a set of if statements, but I want it by a BY group. Eg. Data Want;  set have; by Store Product; if lag(Price) > 35 then Score = 'Bad'; else Score = 'Good'; run; I want this done by each Store and Product based on a time series, I want to look at the Score in previous month, but can't seem to separete the data by Store and Porducts. ... View more

Great info .. thanks PGStats ... View more

To make things interesting, I also need to deal with hills... so sharp dips in the data .. exactly the same pattern, but inverted as above... ... View more

Data is monthly based,  on a 6-month rolling average... All 4 samples have spikes in different ways that I'd like to get rid of..  Thanks Udo Date           Price 2003.03      571 2003.04      483 2003.05      571 2003.06      428 2003.07      428 2003.08      395 2003.09      334 2003.10      242 2003.11      243 2003.12      251 2004.01      267 2004.02      244 ... it could gradually go up to Price= 600 which is fine... And another Sample is Date           Price 2002.11      267 2002.12      278 2003.01      330 2003.02      390 2003.03      571 2003.04      483 2003.05      571 2003.06      428 2003.07      428 2003.08      395 2003.09      334 2003.10      242 2003.11      243 2003.12      251 2004.01      267 2004.02      244 SAMPLE 3 Date           Price 2002.11      267 2002.12      278 2003.01      330 2003.02      390 2003.03      571 2003.04      483 2003.05      571 2003.06      428 2003.07      428 2003.08      395 2003.09      334 2003.10      242 2003.11      243 2003.12      600 2004.01      644 2004.02      674 And Last sample: Date           Price 2002.11      675 2002.12      680 2003.01      688 2003.02      290 2003.03      300 2003.04      323 2003.05      364 2003.06      375 2003.07      380 2003.08      395 2003.09      334 2003.10      242 2003.11      243 2003.12      251 2004.01      267 2004.02      244 ... View more

It's monthly based, but it is on a 6-month rolling average... but where there is a low count, the spikes happen... here's some examples... Date           Price 2003.03      571 2003.04      483 2003.05      571 2003.06      428 2003.07      428 2003.08      395 2003.09      334 2003.10      242 2003.11      243 2003.12      251 2004.01      267 2004.02      244 ... it could gradually go up to Price= 600 which is fine... And another Sample is Date           Price 2002.11      267 2002.12      278 2003.01      330 2003.02      390 2003.03      571 2003.04      483 2003.05      571 2003.06      428 2003.07      428 2003.08      395 2003.09      334 2003.10      242 2003.11      243 2003.12      251 2004.01      267 2004.02      244 ... View more

Thanks Haikuo, but I've tried that before.. the problem is I'm looking for spikes that could be 2 months long or 7 months long, and the data is already seasonal and averaged out... it's these certain spikes that may pop up in certain months and in certain series.  As I need to identify the 'Bad ares' first, and then I can apply the proc expand to them. ... View more

Thanks Udo... i started playing with Lags and differences and am seeing couple of roadblocks in that.. for example if the data starts high and then drops 100% to a normal level, or if the data starts normal and then has a long period increase like more than 7 months, and then back to normla, I still get a spike... too many tricks to cover.. ... View more

Thanks Ballardw, but I mean more like instead of showing 2000.01 ... 2000.02 ..... 2000.03 and so on, which crams it up.. I want to show only say years or quarters or something like 2000.01... 2000.06   that sort of thing. ... View more

Thanks Cynthia, would you know how I can I limit the number of inputs shown on the x-axis (eg. Date..) I have way too many showing ? ... View more

Thanks Astounding.. ... View more

Thank you, Sir.. ... View more

Thanks Art, it works exactly how I envision it.. perfect.. Is there a way I can also see the original price...?   ... View more

Thanks Art,  this looks good.  But I do have one question, does it only apply if there are 3 outliers as per my example?  If in some cases I have 5 or 10 outliers (depending on the BY statement).. I do have Stores and Products series that I have to consider...  I guess I can always extend my Mean.. ... View more