將字串轉換為<K:date,V:int>型別的字典

我正在研究需要來自 BLE 設備的 RSSI 值的 ML 模型。對于這種情況，我創建了一個 Mac 應用程式，我將一個型別的字典存盤<K: Date,v: Int>在文本檔案中。請參閱下面的示例。

string = '[2021-10-17 06:52:00  0000: -47, 2021-10-17 06:52:04  0000: -50, 2021-10-17 06:52:03  0000: -50, 2021-10-17 06:52:02  0000: -47, 2021-10-17 06:52:08  0000: -46, 2021-10-17 06:51:57  0000: -50, 2021-10-17 06:52:09  0000: -48, 2021-10-17 06:52:05  0000: -49, 2021-10-17 06:52:01  0000: -48, 2021-10-17 06:51:58  0000: -50, 2021-10-17 06:51:59  0000: -50, 2021-10-17 06:52:06  0000: -47, 2021-10-17 06:52:07  0000: -48]'

這里，樣本中的負值是 RSSI 值。例如對于前 2 個索引

對于我執行任何計算，我需要資料<Date, Int>在 python 上具有等效型別。如何將上述字串轉換為 Pandas Dataframe 以執行計算？希望這提供了足夠的資訊。先感謝您。

uj5u.com熱心網友回復：

您可以使用re.findall一個小的正則運算式：

作為資料框

string = '[2021-10-17 06:52:00  0000: -47, 2021-10-17 06:52:04  0000: -50, 2021-10-17 06:52:03  0000: -50, 2021-10-17 06:52:02  0000: -47, 2021-10-17 06:52:08  0000: -46, 2021-10-17 06:51:57  0000: -50, 2021-10-17 06:52:09  0000: -48, 2021-10-17 06:52:05  0000: -49, 2021-10-17 06:52:01  0000: -48, 2021-10-17 06:51:58  0000: -50, 2021-10-17 06:51:59  0000: -50, 2021-10-17 06:52:06  0000: -47, 2021-10-17 06:52:07  0000: -48]'

import re
import pandas as pd

df = (pd.DataFrame.from_records(re.findall('([^,] ): (-?\d )(?:, )?', string[1:-1]),
                                columns=['Date', 'RSSI'])
        .astype({'Date': 'datetime64', 'RSSI': 'int'})
      )

輸出：

                         Date RSSI
0   2021-10-17 06:52:00  0000  -47
1   2021-10-17 06:52:04  0000  -50
2   2021-10-17 06:52:03  0000  -50
3   2021-10-17 06:52:02  0000  -47
4   2021-10-17 06:52:08  0000  -46
5   2021-10-17 06:51:57  0000  -50
6   2021-10-17 06:52:09  0000  -48
7   2021-10-17 06:52:05  0000  -49
8   2021-10-17 06:52:01  0000  -48
9   2021-10-17 06:51:58  0000  -50
10  2021-10-17 06:51:59  0000  -50
11  2021-10-17 06:52:06  0000  -47
12  2021-10-17 06:52:07  0000  -48

作為字典

import re
dict(re.findall('([^,] ): (-?\d )(?:, )?', string[1:-1]))

輸出：

{'2021-10-17 06:52:00  0000': '-47',
 '2021-10-17 06:52:04  0000': '-50',
 '2021-10-17 06:52:03  0000': '-50',
 '2021-10-17 06:52:02  0000': '-47',
 '2021-10-17 06:52:08  0000': '-46',
 '2021-10-17 06:51:57  0000': '-50',
 '2021-10-17 06:52:09  0000': '-48',
 '2021-10-17 06:52:05  0000': '-49',
 '2021-10-17 06:52:01  0000': '-48',
 '2021-10-17 06:51:58  0000': '-50',
 '2021-10-17 06:51:59  0000': '-50',
 '2021-10-17 06:52:06  0000': '-47',
 '2021-10-17 06:52:07  0000': '-48'}

作為具有正確型別的字典：

import re
import pandas as pd
{pd.to_datetime(k): int(v)
 for k,v in re.findall('([^,] ): (-?\d )(?:, )?', string[1:-1])}

輸出：

{Timestamp('2021-10-17 06:52:00 0000', tz='UTC'): -47,
 Timestamp('2021-10-17 06:52:04 0000', tz='UTC'): -50,
 Timestamp('2021-10-17 06:52:03 0000', tz='UTC'): -50,
 Timestamp('2021-10-17 06:52:02 0000', tz='UTC'): -47,
 Timestamp('2021-10-17 06:52:08 0000', tz='UTC'): -46,
 Timestamp('2021-10-17 06:51:57 0000', tz='UTC'): -50,
 Timestamp('2021-10-17 06:52:09 0000', tz='UTC'): -48,
 Timestamp('2021-10-17 06:52:05 0000', tz='UTC'): -49,
 Timestamp('2021-10-17 06:52:01 0000', tz='UTC'): -48,
 Timestamp('2021-10-17 06:51:58 0000', tz='UTC'): -50,
 Timestamp('2021-10-17 06:51:59 0000', tz='UTC'): -50,
 Timestamp('2021-10-17 06:52:06 0000', tz='UTC'): -47,
 Timestamp('2021-10-17 06:52:07 0000', tz='UTC'): -48}

uj5u.com熱心網友回復：

你可以這樣做：

In [98]: l = string[1: -1].split(',')
In [140]: d = {i.split(': ')[0]: i.split(': ')[1] for i in l}

In [131]: df = pd.DataFrame(d.items(), columns=['Date', 'RSSI'])
In [132]: df
Out[132]: 
                          Date RSSI
0    2021-10-17 06:52:00  0000  -47
1    2021-10-17 06:52:04  0000  -50
2    2021-10-17 06:52:03  0000  -50
3    2021-10-17 06:52:02  0000  -47
4    2021-10-17 06:52:08  0000  -46
5    2021-10-17 06:51:57  0000  -50
6    2021-10-17 06:52:09  0000  -48
7    2021-10-17 06:52:05  0000  -49
8    2021-10-17 06:52:01  0000  -48
9    2021-10-17 06:51:58  0000  -50
10   2021-10-17 06:51:59  0000  -50
11   2021-10-17 06:52:06  0000  -47
12   2021-10-17 06:52:07  0000  -48

標籤：Python 熊猫 细绳

上一篇：為什么設定物件拆分字串？

下一篇：帶有flto的RHEL8上的C 字串段錯誤（但不是RHEL7）