Extracting 3 largest values?

I just tried it, and I verified that it is picking the correct points using my code example I posted above as well.

PRS1 = point 8

PRS2 = point 1

My array says max order is 8, 1, 6, 2, 5, 4, 7, 3

I

'm really riding the struggle bus. I show the same thing, that 1 should be PRS2. When I look at my actual data/results, PRS2 should be 7, and PRS3 will be 6. I included a screen shot of the values that are computed for PRS1, this is accurate of my part.

If you're trying to hover over the 3 PR variables to see the values, that will confuse you because the variables within the Do/Unitl blocks will only hold the "real" values during execution.

Not specifically related to your request, but I thought it would be helpful as i see your example doesn't have the origin at zero... There's a "small arc" method you should be using to get better accuracy on your results.

Measure the arc with a "Fixed radius" of the nominal value, 3-5 hits should suffice.

align/origin to that "fixed radius" arc segment.

then remeasure the arc segment with your 9 hits, using least squares math.

finally, you can extract your 3 highest points... but after doing this, my guess is that you won't need to (as the 9 hit radius will be more accurate than the 3 extracted hits)

Hey Neil, is there any chance you guys could add .PR as a valid expression extension for .HIT? Unless there is a trick I don't know about to report the ACTUAL polar radius without making extra variables.

I know there's the mathematical shortcuts, and extreme points available, but I think it would be nice to have. 

I know we can report the location of a feature and get the PR, but I believe the command is something like LOC1.PR.MEAS, but only works if it's part of a dimension.

Example, if I did 4CR.HIT[5].PR it just shows 0.

I report the Polar Radius of a single hit 5 and it works just fine

I were to edit the ".HIT[5..5]" to ".HIT[5]" to trim the extra ..5 that PC-DMIS adds when you click a point in the edit window it says invalid index over and over.

ASSIGN/V1=4CR.HIT[5].PR
DIM LOC1= LOCATION OF SET 4CR.HIT[5..5]  UNITS=IN ,$
GRAPH=OFF  TEXT=OFF  MULT=10.00  OUTPUT=BOTH  HALF ANGLE=NO
AX       MEAS    NOMINAL       +TOL       -TOL        DEV     OUTTOL
PR      0.9386     0.9436     0.0000     0.0000    -0.0050     0.0050 <--------
END OF DIMENSION LOC1

WilliamL said:

Hey Neil, is there any chance you guys could add .PR as a valid expression extension for .HIT? Unless there is a trick I don't know about to report the ACTUAL polar radius without making extra variables.

We have very limited resources available and I don't think there would be enough business justification to make this a high priority - so probably not.

 I were to edit the ".HIT[5..5]" to ".HIT[5]" to trim the extra ..5 that PC-DMIS adds when you click a point in the edit window it says invalid index over and over.

That's how it's supposed to work.  You need to use the .HIT[#..#] syntax.

If I did 4CR.HIT[5].XYZ it gives me the XYZ of that hit. Is this only because of dimensioning  that it requires a beginning and end index? That actually makes sense, thanks.

Generally it's better to use the full syntax everywhere.  There are a few fringe cases (like the one you've found) where omitting the .. may work but I don't know them all and it's not reliable.

That seems horribly overcomplicated. I think it would be easier to avoid loops and based on the original post may be more than he needed. See below.

CIR1       =FEAT/CONTACT/CIRCLE/DEFAULT,CARTESIAN,IN,LEAST_SQR
            THEO/<8.5,0,-0.7192>,<0,0,1>,0.375
            ACTL/<8.25,0,-0.7192>,<0,0,1>,0.375
            TARG/<8.25,0,-0.7192>,<0,0,1>
            START ANG=0,END ANG=360
            ANGLE VEC=<1,0,0>
            DIRECTION=CCW
            SHOW FEATURE PARAMETERS=NO
            SHOW CONTACT PARAMETERS=NO
            ASSIGN/V1=FUNCTION((X,Z),SQRT((X^2)+(Z^2)))
            ASSIGN/V2=MAXINDICES(V1(CIR1.HITS.X,CIR1.HITS.Z))
CIR2       =FEAT/CIRCLE,CARTESIAN,IN,LEAST_SQR,NO
            THEO/<4.1977,0,-0.3597>,<0.085367,0,0.9963496>,8.4261
            ACTL/<4.1977,0,-0.3597>,<0.085367,0,0.9963496>,8.4261
            CONSTR/CIRCLE,BF,3D,CIR1.HIT[V2[1]],CIR1.HIT[V2[2]],CIR1.HIT[V2[3]],,
            OUTLIER_REMOVAL/OFF,3
            FILTER/OFF,UPR=0

or do the same but use POLAR coords for the feature and .PA in the maxindices.  No need for the function/sqrt.