In the table on the Image below, ippt_allsi represents precipitation amounts for a given storm. The rest of the fields represent the precipitation amounts for that number of years recurrence intervals. 2, 5,10, 25, 50 and so on. I need a script that will take the precipitation amount for the storm, (appt_allsi) then find the bounding larger and smaller values and interpolate the recurrence year by where the precipitation amount falls between those years values. For instance, for the first record, 3.84 inches of rain fell. The 100 year amount is 3.615 and the 200 year amount is 4.059. How would I return the interpolated recurrence year for the 3.84 inches? I am fairly new to python and am struggling. Could somebody please get me started in the right direction?
Accepted Solutions
absolutely not!!! that would defeat the purpose of using arrays structures and methods in the first place ... Are you using the same Time, Val relationship for all cases of recurrence intervals? If so, you will notice that I fed 3 numbers into the "curve" and it "interpolated 3 values. If you had 1, you would feed it one.
I will write this verbosely, so you can see what can be done. In this case, I will get RI's for the same 3 numbers, but using 2 different T-Val relationships. I could just as easily use one relationship and determine all the interpolations for any number of inputs. Ergo... that is why I was interested in seeing your data construct so I can show you the easiest way of forumulating the array without having to specify field names and datatype during its construct. But anyway, this will give you an idea... and a lesson 
import numpy as np
np.set_printoptions(edgeitems=3,linewidth=80,precision=2,suppress=True,threshold=20, formatter={'float': '{: 0.3f}'.format})
#
t = [5,10,25,50,100,200,500,1000] # fixed time increments
v = [1, 3, 5, 10, 20, 50, 100, 200] # a row in your table of precip
v1 = [2,4,6,11,25,50,140,225]
#
a = np.array(list(zip(t,v,v1)), dtype=[('Time','int32'),('Val_0','float64'), ('Val_1','float64')])
print("array \n{}\ndtype {}\nShape {}".format(a, a.dtype, a.shape))
print("viewed differently\n{}".format(a.reshape((-1,1))))
#
obs = [2, 7, 10]
rec_int = [ np.interp(obs,a[ i ],a['Time']) for i in ['Val_0','Val_1'] ]
print("Interpolation results for {}\ninterval ...{}".format(obs,rec_int))array
[(5, 1.0, 2.0) (10, 3.0, 4.0) (25, 5.0, 6.0) (50, 10.0, 11.0) (100, 20.0, 25.0)
(200, 50.0, 50.0) (500, 100.0, 140.0) (1000, 200.0, 225.0)]
dtype [('Time', '<i4'), ('Val_0', '<f8'), ('Val_1', '<f8')]
Shape (8,)
viewed differently
[[(5, 1.0, 2.0)]
[(10, 3.0, 4.0)]
[(25, 5.0, 6.0)]
[(50, 10.0, 11.0)]
[(100, 20.0, 25.0)]
[(200, 50.0, 50.0)]
[(500, 100.0, 140.0)]
[(1000, 200.0, 225.0)]]
Interpolation results for [2, 7, 10]
interval ...[array([ 7.500, 35.000, 50.000]), array([ 5.000, 30.000, 45.000])]And there will be naysayers ... right Darren 
I get your example when you manually put in the V values. However, the V value and the obs value will need to be dynamic variables. Here is a breakdown of exactly what I am trying to do. If you look at the above Image, its the attribute field of my point feature class. The Precip_Amount is the observed depth. "obs" variable. The next 8 fields are precipitation Frequency depth amounts by year. Each field representing a year amount. 5, 10, 25 and so on. "v" variable. I want to use those and the harcoded "Years" variable to interpolate the ARI field with the recurrence interval. This will have to be done for 370k records with different v and obs values for every record. Then I want to export the points to a raster. Which I know how to do but if I go to an array I need to be able to get back to points.
