Python csv calculate percentage by group

[larafaelivrin]

I have a table (csv file) with three columns:

Wood [m2] Polygon Area [m2]
15 A 50
10 A 50
12 B 30
10 C 30
05 D 50
10 D 50

My aim is to calculate the percentage of wood for each Polygon. I want to print this result into a new csv table:

Polygon Percentage of Wood (%)
A 0.5 (=25/50)
B
C
D

I usually use Python through ArcGIS (arcpy module) but the modules are very slow for certain things. This is why I want to try to solve the question without this module. But I cannot figure out how to do this. Any help is greatly appreciated.

[bvdet]

You would start by opening the file, reading the file, breaking up the file contents to individual parts and saving in a container object such as a list or dictionary, iterate on the container and perform your calculations, print the output or save to disk. Would not you have to do those steps in ArcGIS?

[larafaelivrin]

no, there are arcpy tools which you can call and as I understand they simplify the steps. But the problem is that some of them take very long to run. This website shows me how to read a csv file (https://docs.python.org/2/library/csv.html) and I managed to do that but how can I group the variables? Is there a function?

[bvdet]

Here's an example of manipulating the data after the file is read:

data = """Wood [m2],Polygon,Area [m2]
15,A,50
10,A,50
12,B,30
10,C,30
05,D,50
10,D,50"""
 
dataLines = data.split("\n")
for line in dataLines[1:]:
    items = line.split(",")
    print ("Polygon %s: \nPercentage: %0.0f%%" %
           (items[1], float(items[0])/float(items[2])*100))
    print "========================"

And the output:

>>> Polygon A: 
Percentage: 30%
========================
Polygon A: 
Percentage: 20%
========================
Polygon B: 
Percentage: 40%
========================
Polygon C: 
Percentage: 33%
========================
Polygon D: 
Percentage: 10%
========================
Polygon D: 
Percentage: 20%
========================
>>> 

[larafaelivrin]

ok but with this solution I get several output for Polygon A and D. I am interested in summarizing the wooden Areas for each Polygon which has the same name. For Polygon A for example this would be 15+20/50. Is the quickest way to sum up the outputs or to do this step beforehand? Thanks a lot!!

[bvdet]

I don't think you want 15+(20/50)(operator precedence). I think you want (15+20)/50.

Here's where a dictionary comes in handy:

data = """Wood [m2],Polygon,Area [m2]
15,A,50
10,A,50
12,B,30
10,C,30
05,D,50
10,D,50"""
 
dataLines = data.split("\n")
dd = {}
for line in dataLines[1:]:
    items = line.split(",")
    dd.setdefault(items[1], []).append((float(items[0]), float(items[2])))
 
keys = sorted(dd.keys())
for key in keys:
    print ("Polygon %s: \nPercentage: %0.0f%%" %
           (key, sum((item[0] for item in dd[key]))/dd[key][0][1]*100))
    print "========================"

[larafaelivrin]

I just copied your code and it works perfectly! Thank you so much!! I will try to understand what you did and maybe I can get back to you in case I do not understand something. Thanks!:)

[larafaelivrin]

Another question (sry...): If I import my csv file I get the fallowing structure:

['15', 'A', '50']
['10', 'A', '50']
['12', 'B', '30']
['10', 'C', '30']
['5', 'D', '50']
['10', 'D', '50']

How do you import your csv file without listing each row separately? I don´t seem to be able to figure out what I am doing wrong...

[larafaelivrin]

Aha, maybe I figured out how to do it:

data = open("Test.csv", "r")
print data.read()

but now I get this error:
Traceback (most recent call last):
File "/home/katharina/Desktop/Test.py", line 14, in <module>
dataLines = data.split("\n")
AttributeError: 'file' object has no attribute 'split'

and if I uncomment the dataLines line the fallowing error appears: Traceback (most recent call last):
File "/home/katharina/Desktop/Test.py", line 16, in <module>
for line in data[1:]:
TypeError: 'file' object has no attribute '__getitem__'

Any clue what I am doing wrong?

[bvdet]

There are several ways of doing this. You don't have to create a file object.

data = open("Test.csv", "r").read()

OR

dataLines = [item.strip() for item in open("Test.csv", "r").readlines()