Python 上没有任何库的 CSV 计算

Question

我有这样的 CSV 文件：

ID,TASK1,TASK2,QUIZ1,QUIZ2
11061,50,75,50,78
11062,70,80,60,50
11063,60,75,77,79
11064,52,85,50,80
11065,70,85,50,80

如何获得新行 NO 并在 TASK1,TASK2,QUIZ1,QUIZ2 上获得最高和最低分数。并覆盖它？我限制在 Python 上使用任何库，我的 Expect 输出是：

NO,ID,TASK1,TASK2,QUIZ1,QUIZ2
1,11061,50,75,50,78
2,11062,70,80,60,50
3,11063,60,75,77,79
4,11064,52,85,50,80
5,11065,70,85,50,80
MAX, ,70,85,77,80
MIN, ,50,75,50,50

Answer 1

您可以在r+ 模式下打开文件，这将允许您在不重新打开的情况下读取和写入文件。

简短的（理解起来有点复杂）版本的代码（下面有解释）：

with open(r"filename.csv", "r+") as f:
    next(f)  # skip first line
    [*columns] = zip(*((int(i) for i in line.split(",")[1:]) for line in f))
    print("MAX", "", *map(max, columns), sep=",", file=f)
    print("MIN", "", *map(min, columns), sep=",", file=f)

现在让我们解释一下它是如何工作的。为了获得每列的最大值/最小值，我们必须将列的所有值分组到一个容器中。为了得到行，我们应该用逗号分割每一行并将每个值转换为整数。由于我们不需要第一列的值，我们可以跳过它。所有这些工作都由一行代码完成：

[*columns] = zip(*((int(i) for i in line.split(",")[1:]) for line in f))

同样的，我们可以用更明确的方式来写：

rows = []
for line in f:
    row = []
    for column in line.split(",")[1:]:  # skip first column
        row.append(int(column))  # convert string to int
    rows.append(row)
columns = list(zip(*rows))

[*columns] = iterable 与columns = list(iterable) 相同。

在我们得到包含每列值的元组列表后，我们可以将max() 和min() 应用于每个元组以获得最大值和最小值。我们可以使用生成器表达式来做到这一点：

(max(column) for column in columns)
(min(column) for column in columns)

或者使用map():

map(max, columns)
map(min, columns)

要将结果写入文件，我们需要将值转换为字符串并使用 str.join() 将它们连接回单个字符串：

max_row = ["MAX", ""]  # empty string needed to get double comas
min_row = ["MIN", ""]
for column in columns:
    max_row.append(str(max(column)))
    min_row.append(str(min(column)))
f.write("\n" + ",".join(max_row))
f.write("\n" + ",".join(min_row))

或者，我们可以使用print()。我们可以unpack 打印我们的结果，将"," 传递给sep 参数，这将使python 使用逗号连接所有参数，并将我们的文件对象传递给file 参数以使其写入文件而不是控制台。

print("MAX", "", *map(max, columns), sep=",", file=f)
print("MIN", "", *map(min, columns), sep=",", file=f)

所以这是答案顶部的无聊版本的代码：

with open(r"filename.csv", "r+") as f:
    next(f)  # skip first line
    rows = []
    for line in f:
        row = []
        for column in line.split(",")[1:]:  # skip first column
            row.append(int(column))  # convert string to int
        rows.append(row)
    max_row = ["MAX", ""]  # empty string needed to get double comas
    min_row = ["MIN", ""]
    for column in zip(*rows):
        max_row.append(str(max(column)))
        min_row.append(str(min(column)))
    f.write("\n" + ",".join(max_row))
    f.write("\n" + ",".join(min_row))

---

Upd. 如果要将结果写入单独的文件并添加索引列，您还需要打开一个目标文件进行写入，将源文件的每一行重写到目标（添加它的索引）并将最小值和最大值附加到文件末尾。

修改后的短版代码（变得更加复杂）：

with open(r"input.csv") as inp_f, \
        open(r"output.csv", "w") as out_f:
    out_f.write("NO," + next(inp_f)) 
    [*columns] = zip(*(out_f.write(f"{index},{line.rstrip()}\n") and
                       (int(i) for i in line.split(",")[1:])
                       for index, line in enumerate(inp_f)))
    print("MAX", "", *map(max, columns), sep=",", file=out_f)
    print("MIN", "", *map(min, columns), sep=",", file=out_f)

修改无聊版代码：

with open(r"input.csv") as inp_f, \
        open(r"output.csv", "w") as out_f:
    out_f.write("NO," + next(inp_f))
    rows = []
    for index, line in enumerate(inp_f, 1):
        out_f.write(str(index) + "," + line)
        row = []
        for column in line.split(",")[1:]:  # skip first column
            row.append(int(column))  # convert string to int
        rows.append(row)
    max_row = ["MAX", ""]  # empty string needed to get double comas
    min_row = ["MIN", ""]
    for column in zip(*rows):
        max_row.append(str(max(column)))
        min_row.append(str(min(column)))
    out_f.write("\n" + ",".join(max_row))
    out_f.write("\n" + ",".join(min_row))

---

更新。我做了一些测试(code)。测试结果（越低越好）：

olvin1: 0.687356842
olvin2: 0.5448804249999999
nikeros: 0.7540002289999999
balderman: 0.6384034139999999

因此，此答案中的“无聊”方法展示了最佳性能。

Answer 2

见下文 - 零进口

TASK1 = 1
TASK2 = 2
QUIZ1 = 3
QUIZ2 = 4

MIN = 0
MAX = 1
values = [[None,None],[None,None],[None,None],[None,None]]
lines = []
with open('in.txt') as f:
  for idx,line in enumerate(f):
    if idx > 0:
      lines.append(line)
      parts = line.split(',')
      for i in [TASK1,TASK2,QUIZ1,QUIZ2]:
        if values[i-1][MIN] is None or values[i-1][MIN] > parts[i]:
          values[i-1][MIN] = parts[i].strip()    
        if values[i-1][MAX] is None or values[i-1][MAX] < parts[i]:
          values[i-1][MAX] = parts[i].strip()
with open('out.txt','w') as f: 
  f.write('NO,ID,TASK1,TASK2,QUIZ1,QUIZ2\n') 
  for idx,line in enumerate(lines,1):
    f.write(f'{idx},{line}') 
  f.write('\n')  
  f.write('MAX, ,' + ','.join(str(x[MAX]) for x in values)) 
  f.write('\n')  
  f.write('MIN, ,' + ','.join(str(x[MIN]) for x in values)) 

out.txt

NO,ID,TASK1,TASK2,QUIZ1,QUIZ2
1,11061,50,75,50,78
2,11062,70,80,60,50
3,11063,60,75,77,79
4,11064,52,85,50,80
5,11065,70,85,50,80
MAX, ,70,85,77,80
MIN, ,50,75,50,50

Answer 3

假设您的输入已经在 csv 文件中，您可以使用以下代码：

# This reads in your file
def load_csv(p):
    with open(p, "r") as f:
        return [l.strip().split(",") for l in f]

# This isolates the numeric columns you are interested in
def to_numeric(lines):
    return [list(map(int,line[1:])) for i, line in enumerate(lines) if i > 0]

# This applies a function to the columns results of to_numeric
def apply_function(numeric_lists, func=max):
    return [func(x) for x in zip(*numeric_lists)] 

# This applies the defined function, printing your result in the desired format
def main(p):
    lines = load_csv(p)
    numeric_lists = to_numeric(lines)
    mins = apply_function(numeric_lists, func=min)
    maxs = apply_function(numeric_lists)
    for i, l in enumerate(lines):
        print(("No," if i == 0 else str(i)) + ",".join(l))
    print(f'MAX, ,{",".join(map(str,maxs))}')
    print(f'MIX, ,{",".join(map(str,mins))}')

# Simply call
main(YOUR_FILE_PATH)