Python pandas dataframe - 找到大于特定值的第一个匹配项

Question

I have a dataframe that looks like this:

print(df.head(20))
      date_final  Close*  year
18399 1949-08-08   15.51  1949
18398 1949-08-09   15.37  1949
18397 1949-08-10   15.44  1949
18396 1949-08-11   15.39  1949
18395 1949-08-12   15.32  1949
18394 1949-08-15   15.25  1949
18393 1949-08-16   15.29  1949
18392 1949-08-17   15.46  1949
18391 1949-08-18   15.50  1949
18390 1949-08-19   15.41  1949
18389 1949-08-22   15.37  1949
18388 1949-08-23   15.17  1949
18387 1949-08-24   15.18  1949
18386 1949-08-25   15.22  1949
18385 1949-08-26   15.28  1949
18384 1949-08-29   15.12  1949
18383 1949-08-30   15.21  1949
18382 1949-08-31   15.22  1949
18381 1949-09-01   15.31  1949
18380 1949-09-02   15.29  1949

I have daily data for many years.

I want to find the min and the max (max1) per year. Then I want to find when I had a new max, ie when max2 becomes max1. I want to find the row that this happens

I am able to find the max and the min using the following code

df_1974.loc[[df_1974["Close*"].idxmax()]]['date_final']
df_1974.loc[[df_1974["Close*"].idxmax()]]['Close*']

but I need some help to proceed. Thanks.

Answer 1

Might suggest creating DataFrame with the Max Values, then finding the due max entries:

df = pd.DataFrame(
    data=(
        ('1949-08-08', 15.51),
        ('1949-08-09', 15.37),
        ('1949-08-10', 15.44),
        ('1949-08-11', 15.39),
        ('1949-08-12', 15.32),
        ('1949-08-15', 15.25),
        ('1949-08-16', 15.29),
        ('1949-08-17', 15.46),
        ('1949-08-18', 15.5),
        ('1949-08-19', 15.41),
        ('1949-08-22', 15.37),
        ('1949-08-23', 15.17),
        ('1949-08-24', 15.18),
        ('1949-08-25', 15.22),
        ('1949-08-26', 15.28),
        ('1949-08-29', 15.12),
        ('1949-08-30', 15.21),
        ('1949-08-31', 15.22),
        ('1949-09-01', 15.31),
        ('1949-09-02', 15.29),
    ),
    columns=('date_final', 'Close*'),
    index=range(18399, 18379, -1),
)
df.set_index(pd.to_datetime(df.iloc[:, 0], format='%Y-%m-%d'), inplace=True)
df.drop('date_final', axis=1, inplace=True)
maxes = df.groupby(df.index.year).max()
for _ in range(maxes.shape[0]):
    print(df[np.logical_and(
        df.index.year == maxes.index[_],
        df.iloc[:, 0] == maxes.iloc[_, 0])])

You may not use print but to use pd.concat to retrieve the resulting DataFrame . Hope that helps.