Check if there are "almost matching" points for the 3D coordinate data group.

Introduction

While searching for a better way .., I will summarize it in this way for the time being. ** (If you know a better way, please let me know !!) **

Thing you want to do

① Basic As shown in the figure below, there is Excel that stores 3D coordinate points, and I want to link numbers with the same coordinates. In the example below, the coordinates of S / N-2 and the coordinates of S / N-3 (highlighted in light blue) are exactly the same, so they are judged to be "the same coordinates!".

② Application As the specification of the target I am working on this time, it is difficult for the coordinates to match exactly. Therefore, I would like to judge that they match with a certain amount of Margin. For example, the coordinates of S / N-8 and the coordinates of S / N-9 (highlighted in yellow) match ** almost **, so please judge that they are also "same coordinates".

This time, I couldn't think of another good idea for this "application". .. If you want to extract only the exact match, I thought that you should combine the X coordinate, Y coordinate, and Z coordinate as a character string and judge whether they match.

The resulting code looks like this

First, create a function that can be judged including margin. In the following cases, the error ± 100 is ignored, and True is returned if the three coordinates match with margin included.

def coordinates_check(ax,ay,az,bx,by,bz):
    margin=100
    flag = 0
    result=False
    
    if ax-margin < bx < ax+margin:
        flag = flag +1
    if ay-margin < by < ay+margin:
        flag = flag +1
    if az-margin < bz < az+margin:
        flag = flag +1
        
    if flag >2:
        result = True

    return result

Use this function to read the lines of the excel file from above. The start number of the second for statement is devised so that the readings are not duplicated.

df_test = pd.read_excel('sample.xlsx')

for i in range(len(df_test)):
    ax= df_test.loc[i,"X coordinate"]
    ay= df_test.loc[i,"Y coordinate"]
    az= df_test.loc[i,"Z coordinate"]
    k = i+1
    for j in range(k,len(df_test)):
        bx = df_test.loc[j,"X coordinate"]
        by = df_test.loc[j,"Y coordinate"]
        bz = df_test.loc[j,"Z coordinate"]
        if coordinates_check(ax,ay,az,bx,by,bz):
            print("Matched:",df_test.loc[i,"S/N"],"-->",df_test.loc[j,"S/N"])

As a result, it is done well.

I wish I had a small amount of data ..

The data I want to use this time is over 10,000 lines, and it took ** 3 hours ** to execute the above code ... It was around this time that I thought there should be a better way. If you know a good way, please help!