python - 是否可以让 csv.DictReader 将某些列格式化为字符串以外的格式？

Question

csv.DictReader() 默认情况下将列中的值作为字符串提取。 有没有办法指定某些列的转换？

每次访问由 csv.DictReader() 创建的包含非字符串元素的字典列表时，我最终都会做很多这样的小烦恼:

with open("data.csv","r") as data_file:
    items = csv.DictReader(data_file, fieldnames=('id', 'length', 'note'))
    for item in items:
        item['length'] = float(item['length'])  #### <--- MINOR ANNOYANCE 
        # ... do loop stuff

如果我可以告诉 csv.DictReader 当它到达某个字段时，它应该执行 float() (或 int ()、date() 等)转换。

Answer 1

不是开箱即用的，不是。您可以子类化 DictReader() 或创建一个生成器函数来为您映射行:

def convert_fields(iterable, **conversions):
    for item in iterable:
        for key in item.viewkeys() & conversions:
            item[key] = conversions[key](item[key])
        yield item

对于 Python 3，请用 dict.viewkeys() 替换 dict.keys()，因为 Python 3 默认返回字典 View 。

用它包装您的 csv.DictReader() ，为每列添加转换函数:

with open("data.csv","r") as data_file:
    items = csv.DictReader(data_file, fieldnames=('id', 'length', 'note'))
    items = convert_fields(items, length=float)
    for item in items:
        # item['length'] is now always a float

对于未映射到 python 标识符(带空格等)的列名，请使用 **{..} 语法传入字典:

with open("data.csv","r") as data_file:
    items = csv.DictReader(data_file, fieldnames=('id', 'length', 'note'))
    fieldconv = {'id': int, 'length': float, 'spaced column': float}
    items = convert_fields(items, **fieldconv)
    for item in items:
        # item['length'] and item['spaced column'] are now floats
        # item['id'] is always an int