python - 组合 DataFrame 行以消除 NaN

Question

我有一个从数据记录器创建的 DataFrame，其中每个数据点都有自己的时间戳，如下所示:

df_orig = pd.DataFrame(
    {
        "val1": [ 1, np.nan, np.nan, 11, np.nan, np.nan, 21, np.nan, np.nan, ],
        "val2": [ np.nan, 2, np.nan, np.nan, 12, np.nan, np.nan, 22, np.nan, ],
        "val3": [ np.nan, np.nan, 3, np.nan, np.nan, 13, np.nan, np.nan, 23, ],
    },
    index=pd.to_datetime( [
        "2021-01-01 00:00", "2021-01-01 00:00:01", "2021-01-01 00:00:02",
        "2021-01-01 00:01", "2021-01-01 00:01:01", "2021-01-01 00:01:02",
        "2021-01-01 00:02", "2021-01-01 00:02:01", "2021-01-01 00:02:02",
    ] )
)

                     val1  val2  val3
2021-01-01 00:00:00   1.0   NaN   NaN
2021-01-01 00:00:01   NaN   2.0   NaN
2021-01-01 00:00:02   NaN   NaN   3.0
2021-01-01 00:01:00  11.0   NaN   NaN
2021-01-01 00:01:01   NaN  12.0   NaN
2021-01-01 00:01:02   NaN   NaN  13.0
2021-01-01 00:02:00  21.0   NaN   NaN
2021-01-01 00:02:01   NaN  22.0   NaN
2021-01-01 00:02:02   NaN   NaN  23.0

我实际上不需要记录每个数据点的精确度。我想通过消除 NaN 并合并非常接近的行来压缩 DataFrame。结果应该如下所示:

                     val1  val2  val3
2021-01-01 00:00:00     1     2     3
2021-01-01 00:01:00    11    12    13
2021-01-01 00:02:00    21    22    23

有办法做到这一点吗？

Answer 1

如果可能，使用 max 或 min 或 first 使用每分钟重新采样简化解决方案:

df = df_orig.resample('Min').max()
print (df)
                     val1  val2  val3
2021-01-01 00:00:00   1.0   2.0   3.0
2021-01-01 00:01:00  11.0  12.0  13.0
2021-01-01 00:02:00  21.0  22.0  23.0