Compare attribute columns of a feature class in ArcGIS

Question

Is there a solution in ArcGIS 10x? Thanks. I used GME (newer Hawth's Tools) and was able to count the number of pixels of any raster class within a vector polygon. I am trying to map crops within a field, so sometimes there can be different classes (crops) within a field polygon. GME spits out class counts as different columns. So if I have 3 raster classes, three count columns. I am looking for a solution to compare the counts of these 3 columns and identify the highest count and attribute that to a new column. I.e. I want to identify the dominant crop within the field. I can look at the column count numbers and identify the dominant class, but this isn't very efficient when you have 100's of fields. Sothing similar is what I am looking for: http://www.get-digital-help.com/2009/07/17/extract-largest-values-from-two-columns-using-array-formula-in-excel/

---

This still doesn't work and Python gives error. I edited the V#V1 as Canola, V#V2 as Barley, V#V3 as Wheat by adding three new text fields, ![enter image description here][1]and then ran:

GetMaxFieldName({"Canola": !Canola!,"Barley": !Barley!,"Wheat": !Wheat!})

A friend helped me solve it in Excel. 1) Run MAX formula in a new column ("Max"; column O) on the four classes (rows K1:N1) to identify the highest count. 2) Add new column ("Crop", column P) after Max and run: =INDEX($K$1:$N$1,1,MATCH(O2,K2:N2,0))

See image links: http://postimage.org/image/68ham84uv/ http://postimage.org/image/r1sz7e21l/

Let me know if you solve this with Python.

Answer 1

Am I correct to assume that you have multiple rows of counts per class, much like pictured in the Excel sample you provided? And you want highest count per class or simply highest overall count? (You can actually do both, 'subtotal' highest and overall highest.)

You can use a search cursor to read the row pairs and write them (with an insert cursor) to a new table that can then be easily sorted (and ArcMap has multiple field sort functionality). You would do this, say using your 1st row as an example, by reading 'across' your 3 class example for 3 pairs of values, as in:

classVal1, count1, classVal2, count2, classVal3, count3

..and populating that in your insert table as:

(row 1): classVal1, count1, class1
(row 2): classVal2, count2, class2
(row 3): classVal3, count3, class3

Then reading successive rows, following the pattern above, the 3rd column (field) will behave as a 'counter', reflecting what original class the count values came from, if that is relevant info you need....

So, after writing the output table, you can easily sort descending by count to get the highest overall count and corresponding classVal...or sort by count and class to get the highest counts per class and corresponding classVals. Is that what you need?

And, as a further note, you can actually use a sort in conjunction with the search cursor on this result to output the highest values if max value(s) are all you wish as final output, or if you have a preference for the summary statistics or frequency tools, use them instead.

Answer 2

Assuming that each row is one crop field polygon, and you want the maximum value of the three count columns (for each row). You can simply use the Python built-in function max in the Field calculator.

Firstly add a new field, then calculate the maximum value for each record from the list of fields (your count columns):

max([!field1!, !field2!, !field3!])

Other field calculator examples.

To see which field name contains the highest value, try this advice (at the end of the thread)

Hope this can be useful.