根据特定列中的值对整个 .csv 进行排序

Question

我有一个csv包含不同收入的文件。我想按收入从高到低对 csv 文件进行排序。我无法找到如何在不使用 python 的情况下在终端中执行此操作。

我不想使用Python。

我想使用简单的东西，比如mlr// 。sedawk

输入：

name,location,capital,profit-lost,revenue,employees,year
company1,location1,35527.19,-33226.25,,0.70,2020
company2,location2,-155921.70,-146.03,,,2020
company3,location3,1873134.74,778424.56,13320152.32,16.90,2020
company4,location4,1050987.60,426317.61,,24.90,2021
company5,location5,368506.18,11997.04,,,2019
company6,location6,7965648.89,369947.14,64413602.44,103.30,2019
company7,location7,1531534.27,125750.94,3054307.36,12.10,2020
company8,location8,6161574.62,906591.96,124804038.64,51.30,2021

输出：

name,location,capital,profit-lost,revenue,employees,year
company8,location8,6161574.62,906591.96,124804038.64,51.30,2021
company6,location6,7965648.89,369947.14,64413602.44,103.30,2019
company3,location3,1873134.74,778424.56,13320152.32,16.90,2020
company7,location7,1531534.27,125750.94,3054307.36,12.10,2020
company1,location1,35527.19,-33226.25,,0.70,2020
company2,location2,-155921.70,-146.03,,,2020
company4,location4,1050987.60,426317.61,,24.90,2021
company5,location5,368506.18,11997.04,,,2019

收入空到几十亿。

希望有人也能帮助我解决这个问题

Answer 1

使用sort：

cat input.csv | (sed -u 1q; sort -t, -r -n -k5) 

需要sed -u 1q忽略sort标头。它基本上意味着，处理第一行并退出，然后将剩余的传递给sort. -u是 的缩写--unbuffered，告诉sed不要缓冲线路。

排序的标志：

• -t,将分隔符指定为逗号。
• -r使排序后的输出降序排列。默认是升序。
• -n按数字排序。
• -k5对第五个键/列进行排序。

演示：

$ cat input.csv | (sed -u 1q; sort -t, -r -n -k5)
name,location,capital,profit-lost,revenue,employees,year
company8,location8,6161574.62,906591.96,124804038.64,51.30,2021
company6,location6,7965648.89,369947.14,64413602.44,103.30,2019
company3,location3,1873134.74,778424.56,13320152.32,16.90,2020
company7,location7,1531534.27,125750.94,3054307.36,12.10,2020
company5,location5,368506.18,11997.04,,,2019
company4,location4,1050987.60,426317.61,,24.90,2021
company2,location2,-155921.70,-146.03,,,2020
company1,location1,35527.19,-33226.25,,0.70,2020

Answer 2

因此，您希望按数字降序对收入进行（稳定）排序，这听起来在 Miller 中应该很容易，除了它的空处理规则说：

具有一个或多个空排序字段值的记录在具有所有排序字段值的记录之后排序

这意味着它们首先按降序排序：

$ mlr --csv sort -nr revenue file.csv
name,location,capital,profit-lost,revenue,employees,year
company1,location1,35527.19,-33226.25,,0.70,2020
company2,location2,-155921.70,-146.03,,,2020
company4,location4,1050987.60,426317.61,,24.90,2021
company5,location5,368506.18,11997.04,,,2019
company8,location8,6161574.62,906591.96,124804038.64,51.30,2021
company6,location6,7965648.89,369947.14,64413602.44,103.30,2019
company3,location3,1873134.74,778424.56,13320152.32,16.90,2020
company7,location7,1531534.27,125750.94,3054307.36,12.10,2020

然而，使用then 链接，可以直接使用将数字 0 分配给空收入的键进行装饰-排序-取消装饰：

$ mlr --csv put '$key = is_empty($revenue) ? 0 : $revenue' \
    then sort -nr key then cut -x -f key file.csv
name,location,capital,profit-lost,revenue,employees,year
company8,location8,6161574.62,906591.96,124804038.64,51.30,2021
company6,location6,7965648.89,369947.14,64413602.44,103.30,2019
company3,location3,1873134.74,778424.56,13320152.32,16.90,2020
company7,location7,1531534.27,125750.94,3054307.36,12.10,2020
company1,location1,35527.19,-33226.25,,0.70,2020
company2,location2,-155921.70,-146.03,,,2020
company4,location4,1050987.60,426317.61,,24.90,2021
company5,location5,368506.18,11997.04,,,2019

Answer 3

使用所有 Unix 系统上可用的强制 POSIX 工具：

$ { head -n 1; sort -t, -k5,5rn; } < file
name,location,capital,profit-lost,revenue,employees,year
company8,location8,6161574.62,906591.96,124804038.64,51.30,2021
company6,location6,7965648.89,369947.14,64413602.44,103.30,2019
company3,location3,1873134.74,778424.56,13320152.32,16.90,2020
company7,location7,1531534.27,125750.94,3054307.36,12.10,2020
company1,location1,35527.19,-33226.25,,0.70,2020
company2,location2,-155921.70,-146.03,,,2020
company4,location4,1050987.60,426317.61,,24.90,2021
company5,location5,368506.18,11997.04,,,2019

请参阅下面的评论和head 可以读取比输出更多的输入行吗？有关上述脚本的其他重要信息。