In DolphinDB’s Python API, what’s the difference between using s.loadTable and s.table to load an in-CodePudding

t1 = s.loadTable(tableName="test")
t2 = s.table(data='test')

Both lines above load the “test“ table. What’s the difference?

CodePudding user response：

From what i can see from the documentation, s.table uploads a new table (see here) and s.loadTable reads an already existing table form the DB (documentation here)

CodePudding user response：

Both methods load the specified table through constructing a Table object. The differences are as follows:

loadTable doesn’t support garbage collection (the automatic memory management mechanism of the Table class that reclaims the allocated memory no longer in reference) whereas table does.
table also accepts a local pandas DataFrame as input parameter. Here’s an example:

Define function createDemonDataFrame() with the below script to create a pandas DataFrame.

import pandas as pd

def createDemoDataFrame():
    data = {'cid': np.array([1, 2, 3], dtype=np.int32),
            'cbool': np.array([True, False, np.nan], dtype=np.bool),
            'cchar': np.array([1, 2, 3], dtype=np.int8),
            'cshort': np.array([1, 2, 3], dtype=np.int16),
            'cint': np.array([1, 2, 3], dtype=np.int32),
            'clong': np.array([0, 1, 2], dtype=np.int64),
            'cdate': np.array(['2019-02-04', '2019-02-05', ''], dtype='datetime64[D]'),
            'cmonth': np.array(['2019-01', '2019-02', ''], dtype='datetime64[M]'),
            'ctime': np.array(['2019-01-01 15:00:00.706', '2019-01-01 15:30:00.706', ''], dtype='datetime64[ms]'),
            'cminute': np.array(['2019-01-01 15:25', '2019-01-01 15:30', ''], dtype='datetime64[m]'),
            'csecond': np.array(['2019-01-01 15:00:30', '2019-01-01 15:30:33', ''], dtype='datetime64[s]'),
            'cdatetime': np.array(['2019-01-01 15:00:30', '2019-01-02 15:30:33', ''], dtype='datetime64[s]'),
            'ctimestamp': np.array(['2019-01-01 15:00:00.706', '2019-01-01 15:30:00.706', ''], dtype='datetime64[ms]'),
            'cnanotime': np.array(['2019-01-01 15:00:00.80706', '2019-01-01 15:30:00.80706', ''], dtype='datetime64[ns]'),
            'cnanotimestamp': np.array(['2019-01-01 15:00:00.80706', '2019-01-01 15:30:00.80706', ''], dtype='datetime64[ns]'),
            'cfloat': np.array([2.1, 2.658956, np.NaN], dtype=np.float32),
            'cdouble': np.array([0., 47.456213, np.NaN], dtype=np.float64),
            'csymbol': np.array(['A', 'B', '']),
            'cstring': np.array(['abc', 'def', ''])}
    return pd.DataFrame(data)

Call table to upload the DataFrame to the DolphinDB server with the name “testDataFrame“. Call loadTable to view the data of the uploaded table.

dt = s.table(data=createDemoDataFrame(), tableAliasName="testDataFrame")
print(s.loadTable("testDataFrame").toDF())

# output
   cid  cbool  cchar  cshort  cint  ...             cnanotimestamp    cfloat    cdouble csymbol cstring
0    1   True      1       1     1  ... 2019-01-01 15:00:00.807060  2.100000   0.000000       A     abc
1    2  False      2       2     2  ... 2019-01-01 15:30:00.807060  2.658956  47.456213       B     def
2    3   True      3       3     3  ...                        NaT       NaN        NaN