'numpy.float64' object is not iterable error when using an array

I am trying to evaluate an ARIMA model. To evaluate the model, I want to use 5 days into the future as test data.

forecast_data = model_fit.forecast(steps=5)

# Last 5 days of forecast data
pred_arima_y = forecast_data[0].tolist()

# 5 days of actual data
test_y = appl_test_df.price.values

# Minimum forecast data for the last 5 days
pred_y_lower = []
# Maximum forecast data for the last 5 days
pred_y_upper = []

for lower_upper in forecast_data[2]:
    lower = lower_upper[0]
    upper = lower_upper[1]
    pred_y_lower.append(lower)
    pred_y_upper.append(upper)

And I get the following error

TypeError                                 Traceback (most recent call last)
<ipython-input-44-309f07141e72> in <cell line: 16>()
     14 pred_y_upper = []
     15 
---> 16 for sub_array in forecast_data[2]:
     17     for lower_upper in sub_array:
     18         lower = lower_upper[0]

TypeError: 'numpy.float64' object is not iterable

To fix the code, I tried using a nested loop to iterate over the entire 2D array forecast_data[2], or tried other methods, but they all failed.

I added some more code to the previous one. The preceding partial code looks like this. ARIMA was not available, so we switched to SARIMAX halfway through.

from statsmodels.tsa.arima_model import ARIMA
from statsmodels.tsa.statespace.sarimax import SARIMAX
import statsmodels.api as sm

# model = ARIMA(appl_train_df.price.values, order = (2,1,2))
# model_fit = model.fit(trend = 'c', full_output = True, disp = True)

model = SARIMAX(endog = appl_train_df.price.values, order = (2, 1, 2), seasonal_order = (1, 1, 1, 12))
model_fit = model.fit(trend = 'c', full_output = True, disp = True)
# model_fit = model.fit(disp=0)

print(model_fit.summary())

                                      SARIMAX Results                                       
============================================================================================
Dep. Variable:                                    y   No. Observations:                  515
Model:             SARIMAX(2, 1, 2)x(1, 1, [1], 12)   Log Likelihood               -3919.268
Date:                              Tue, 02 Apr 2024   AIC                           7852.536
Time:                                      01:36:55   BIC                           7882.066
Sample:                                           0   HQIC                          7864.122
                                              - 515                                         
Covariance Type:                                opg                                         
==============================================================================
                 coef    std err          z      P>|z|      [0.025      0.975]
------------------------------------------------------------------------------
ar.L1          1.8483      0.012    154.720      0.000       1.825       1.872
ar.L2         -0.9506      0.011    -82.778      0.000      -0.973      -0.928
ma.L1         -1.8954      0.010   -187.159      0.000      -1.915      -1.876
ma.L2          0.9839      0.010     98.924      0.000       0.964       1.003
ar.S.L12       0.1830      0.018     10.002      0.000       0.147       0.219
ma.S.L12      -0.9623      0.032    -30.038      0.000      -1.025      -0.900
sigma2      2.977e+05   1.44e+04     20.616      0.000    2.69e+05    3.26e+05
===================================================================================
Ljung-Box (L1) (Q):                   1.85   Jarque-Bera (JB):               665.47
Prob(Q):                              0.17   Prob(JB):                         0.00
Heteroskedasticity (H):               0.35   Skew:                             0.05
Prob(H) (two-sided):                  0.00   Kurtosis:                         8.64
===================================================================================

And I checked the value with

print(forecast_data[2].shape)

, but the result is

( )

, It is very strange.