问题
I am developing an app which obtains financial data from the Federal Reserve, stores it in a SQlite database using Flask-SQlalchemy, and then serves the data to my front-end dashboard (Bokeh).
The data I am using is structured as follows:
date realtime_end realtime_start value
0 1962-01-02 2018-08-02 2018-08-02 4.06
1 1962-01-03 2018-08-02 2018-08-02 4.03
2 1962-01-04 2018-08-02 2018-08-02 3.99
3 1962-01-05 2018-08-02 2018-08-02 4.02
4 1962-01-08 2018-08-02 2018-08-02 4.03
5 1962-01-09 2018-08-02 2018-08-02 4.05
6 1962-01-10 2018-08-02 2018-08-02 4.07
7 1962-01-11 2018-08-02 2018-08-02 4.08
8 1962-01-12 2018-08-02 2018-08-02 4.08
9 1962-01-15 2018-08-02 2018-08-02 4.10
10 1962-01-16 2018-08-02 2018-08-02 4.13
11 1962-01-17 2018-08-02 2018-08-02 4.12
12 1962-01-18 2018-08-02 2018-08-02 4.11
13 1962-01-19 2018-08-02 2018-08-02 4.11
14 1962-01-22 2018-08-02 2018-08-02 4.09
15 1962-01-23 2018-08-02 2018-08-02 4.11
16 1962-01-24 2018-08-02 2018-08-02 4.10
17 1962-01-25 2018-08-02 2018-08-02 4.11
18 1962-01-26 2018-08-02 2018-08-02 4.11
19 1962-01-29 2018-08-02 2018-08-02 4.12
20 1962-01-30 2018-08-02 2018-08-02 4.11
21 1962-01-31 2018-08-02 2018-08-02 4.10
22 1962-02-01 2018-08-02 2018-08-02 4.09
23 1962-02-02 2018-08-02 2018-08-02 4.08
24 1962-02-05 2018-08-02 2018-08-02 4.07
25 1962-02-06 2018-08-02 2018-08-02 4.06
26 1962-02-07 2018-08-02 2018-08-02 4.07
27 1962-02-08 2018-08-02 2018-08-02 4.07
28 1962-02-09 2018-08-02 2018-08-02 4.05
29 1962-02-12 2018-08-02 2018-08-02 NaN
In my SQLalchemy model, I am trying to store data from the "value" column as follows:
value = db.Column(db.Float)
It is currently in the float64 datatype, but SQLite needs it to be in a generic Python float type (rather than NumPy, which is float64). I am trying to convert the "value" column into generic Python float values as follows:
ValueData = GeneralData['value'].astype(float) #slices the 'value' column from the dataframe
ValueForSQL = Graph(value = ValueData) #stores values in the 'value' column of SQLite table.
I am however getting the following error when running the code: "StatementError cannot convert the series to class 'float'".
The issue appears to be that SQlite cannot process series or lists of data. If I use GeneralData['value'].iloc[3], for example, it works fine. This is because the data is already in float64 format.
SQlite accepts this because it is an individual data point, rather than a list or series. 'Value' is column of thousands of values, however. How can get SQlite to accept the entire series?
回答1:
astype will still be converting it to float64. How about converting the series to a list and mapping it to float values
list(map(float,GeneralData.value))
来源：https://stackoverflow.com/questions/51661061/how-to-insert-timeseries-data-into-sqlite-using-sqlalchemy