I am using python pandas to execute query on MySql. In UI side using Flot API to represent the MySql data. Below is the existing implementation,
query2 = f"""select time, value from NWSTAT
where time > \'{from_date}\'
and time < \'{to_date}\'"""
result2 = pd.read_sql(query2, engine)
return result2.to_json(orient='records')
Getting result in below format
[{"time": 1581931200000, "value": 0.0}, {"time": 1581931200000, "value": 0.0},
{"time": 1581931200000, "value": 0.0}, {"time": 1581931200000, "value": 0.0}]
From this response I am creating belwo structure for Flot API in UI Javascript side,
[[1581931200000,0],[1581931200000,0],[1581931200000,0],[1581931200000,0]]
Is there any way to do this in python side itself with out any iterations? Directly from query result.
Using Flask server. UI side: JQuery, Handlebar JS
EDIT: In accepted answer second approach takes lesser time.. Below is the time taken for both approach for 240k records
First one: --- 1.6689300537109375e-06 seconds ---
Second one: --- 0.5330650806427002 seconds ---
Problem is if convert both columns to numpy array format of integers is changed to floats.
print (json.dumps(result2.to_numpy().tolist()))
First idea is create lists from .values()
of dictionaries and convert to json
:
import json
query2 = f"""select time, value from NWSTAT
where time > \'{from_date}\'
and time < \'{to_date}\'"""
result2 = pd.read_sql(query2, engine)
return json.dumps([list(x.values()) for x in result2.to_dict(orient='records')])
Or change fomrat by DataFrame.to_dict
with l
for lists and then use zip
with mapping lists, last convert to json
:
import json
query2 = f"""select time, value from NWSTAT
where time > \'{from_date}\'
and time < \'{to_date}\'"""
result2 = pd.read_sql(query2, engine)
return json.dumps(list(map(list, zip(*result2.to_dict(orient='l').values()))))