使用Python從多個時間序列創建Pandas Time Serie

Question

我有這樣的多個時間序列：

DATE
2015-10-10 01:00:00    955.0
2015-10-11 01:00:00    702.0
2015-10-12 01:00:00    597.0
2015-10-13 01:00:00    516.0
2015-10-14 01:00:00    554.0

DATE
2015-10-10 02:00:00    972.0
2015-10-11 02:00:00    646.0
2015-10-12 02:00:00    529.0
2015-10-13 02:00:00    554.0
2015-10-14 02:00:00    540.2

DATE
2015-10-10 03:00:00    964.0
2015-10-11 03:00:00    707.0
2015-10-12 03:00:00    557.0
2015-10-13 03:00:00    515.0
2015-10-14 03:00:00    437.2

我想要做的是從這些時間序列中創建一個ordred和唯一的Time Serie來獲得這個結果：

DATE
2015-10-10 00:00:00     622.0
2015-10-10 01:00:00     955.0
2015-10-10 02:00:00     972.0
2015-10-10 03:00:00     964.0
2015-10-10 04:00:00     914.0
...
2015-10-11 00:00:00     923.0
2015-10-11 01:00:00     955.0
2015-10-11 02:00:00     646.0

Answer 1

您可以使用concat逐行連接並在結果上調用sort_index()以獲得所需的結果：

pd.concat(list_of_series).sort_index()

Answer 2

您也可以嘗試使用RedBlackPy 。 該庫旨在有效地處理動態數據（例如，時間序列）。

import redblackpy as rb
from datetime import datetime

series = rb.Series(dtype='float32', interpolate='floor')

# When you insert items, it is automatically sorted.
# Because rb.Series uses red-black trees as a core structure,
# and you can add items and doesn't think about order.
series.insert(datetime(2015,10,10), 955).

# if you have a list of rb.Series objects, than to construct sorted
# union of theirs keys in efficient way you can use rb.SeriesIterator
# which do not use additional memory to concat because it is a generator.
iterator = rb.SeriesIterator(list_of_rb_Series)

for key in iterator('forward'): # or 'reverse' order
    key # key from sorted union of the keys, it is constructed inplace