获取元组列表中浮点数的最大值？

Question

i have the following list:

erra_eus_repo = [(u'RHSA-2017:2796', u'6.7'), (u'RHSA-2017:2796', u'6.8'), (u'RHSA-2017:2794', u'7.2'), (u'RHSA-2017:2793', u'7.3')]

what I am trying to take the floating point numbers from each tuple :

6.7, 6.8 ,7.2, 7.3

and get the max number for each version that before the dot .ie :

new_list = [ 6.8, 7.3 ]

Note that max() will not work here, since if I have 5.9 and 5.11, I will get the max as 5.9, I want the result to be 5.11 since 11 > 9.

What I have tried:

eus_major = []
eus_minor = []
for major in erra_eus_repo:
    minor = (major[1][2])
    major = (major[1][0])
    if major not in eus_major:
            eus_major.append(major)

    if minor not in eus_minor:
            eus_minor.append(minor)
 print(eus_major, eus_minor)

currently i am getting:

[u'6', u'7'] [u'7', u'2', u'3']

Answer 1

You can achieve this for instance with a combination of groupby and sorting:

from itertools import groupby

srt_list = sorted(erra_eus_repo, key=lambda x: x[1]);

max_list = []
for key, group in groupby(srt_list, lambda x: x[1].split('.')[0]):
        max_el = max(list(group), key = lambda y: int(y[1].split('.')[1]))
        max_list.append(float(max_el[1]))

First the array is sorted based on second element of each tuple to get sequences of elements with matching non-decimal number for grouping with groupby . groupby groups the elements into just that - each group will represent a sequence XZ with common X. In each of these sequences - groups the program finds the one with maximum decimal part treated as a stand-along number. The whole number is then appended to the list with max values as a float.

Answer 2

Do not treat the version numbers as floating point, treat as '.' separated strings. Then split each version string (split on '.') and compare. Like this:

def normalize_version(v):
    return tuple(map(int, v.split('.')))

Then you can see:

>>> u'5.11' > u'5.9'
False

>>> normalize_version(u'5.11') > normalize_version(u'5.9')
True

Answer 3

Here is another take on the problem which provides the highest value for each RHSA value (which is want I think you're after):

erra_eus_repo = [(u'RHSA-2017:2796', u'6.7'), (u'RHSA-2017:2796', u'6.8'), (u'RHSA-2017:2794', u'7.2'), (u'RHSA-2017:2793', u'7.3')]

eus_major = {}

for r in erra_eus_repo:

    if r[0] not in eus_major.keys():
        eus_major[r[0]] = 0

    if float(r[1]) > float(eus_major[r[0]]):
        eus_major[r[0]] = r[1]

print(eus_major)

output:

{'RHSA-2017:2796': '6.8', 'RHSA-2017:2794': '7.2', 'RHSA-2017:2793': '7.3'}

I left the value as a string, but it could easily be cast as a float.

Answer 4

The following simply uses the built-in min and max functions:

erra_eus_repo = [(u'RHSA-2017:2796', u'6.7'),
                 (u'RHSA-2017:2796', u'6.8'),
                 (u'RHSA-2017:2794', u'7.2'),
                 (u'RHSA-2017:2793', u'7.3')]

eus_major = max(float(major[1])for major in erra_eus_repo)
eus_minor = min(float(major[1])for major in erra_eus_repo)
newlist = [eus_minor, eus_major]
print(newlist)  # -> [6.7, 7.3]

Answer 5

This may look like you're trying to compare decimal values, but you really aren't. It goes without saying (but I will) that while 9<11, .9>.11. So the idea of splitting the number into two separate values is really the only way to get a valid comparison.

The list is a list of lists - you have the master and each has a sub-list of RHSA and a value. Apparently you want discard the first item in the list and only get the (I assume) version of that item. Here's some code that, while crude, will give you an idea of what to do. (I'd welcome comments on how to clean that up...) So I've taken the lists, split them, then split the versions into major and minor, then compared them and if nothing exists in the list, add the value. I also added for sake of testing a 6.11 version number.

lstVersion = []
lstMaxVersion=[]
erra_eus_repo = [(u'RHSA-2017:2796', u'6.7'), (u'RHSA-2017:2796', u'6.8'), (u'RHSA-2017:2796', u'6.11'), (u'RHSA-2017:2794', u'7.2'), (u'RHSA-2017:2793', u'7.3')]
for strItem in erra_eus_repo:
    lstVersion.append(strItem[1])

for strVersion in lstVersion:
    blnAdded = False
    intMajor = int(strVersion.split('.')[0])
    intMinor = int(strVersion.split('.')[1])
    print 'intMajor: ', intMajor
    print 'intMinor:' , intMinor
    for strMaxItem in lstMaxVersion:
        intMaxMajor = int(strMaxItem.split('.')[0])
        intMaxMinor = int(strMaxItem.split('.')[1])
        print 'strMaxitem: ', strMaxItem
        print 'intMaxMajor: ', intMaxMajor
        print 'intMaxMinor: ', intMaxMinor
        if intMajor == intMaxMajor:
            blnAdded = True
            if intMinor > intMaxMinor:
                lstMaxVersion.remove(strMaxItem)
                lstMaxVersion.append(str(intMajor)+'.'+str(intMinor))
    if not blnAdded:
        lstMaxVersion.append(str(intMajor)+'.'+str(intMinor))