Creating QgsVectorLayer from Pandas dataframe

Question

I am writing a python script for QGIS.

I have a pandas dataframe that I would like to use in order to create a QgsVectorLayer.

At the moment I am saving the data frame into a .csv and then I load it:

name_csv = 'virtuali.csv'    
df.to_csv(name_csv,index=False)
path = "file:///" + name_csv + "?encoding=%s&delimiter=%s&xField=%s&yField=%s&crs=%s" % ("UTF-8",",", "Longitudine", "Latitudine","epsg:4326")
mylayer = QgsVectorLayer(path, "pandas_layer", "delimitedtext")

Is there a way to create the Pandas dataframe directly without saving and reloading it?

Answer 1

Your method is the simplest. However, there is an alternative by creating a new layer in QGIS and completing it by browsing your dataframe pandas.

import pandas as pd 
# Declaration of my pandas dataframe 
d = {'col1' : [1,2], 'col2': [3,4]}
df = pd.DataFrame(data=d)

# Creation of my QgsVectorLayer with no geometry 
temp = QgsVectorLayer("none","result","memory")
temp_data = temp.dataProvider()
# Start of the edition 
temp.startEditing()

# Creation of my fields 
for head in df : 
    myField = QgsField( head, QVariant.Double )
    temp.addAttribute(myField)
# Update     
temp.updateFields()

# Addition of features
# [1] because i don't want the indexes 
for row in df.itertuples():
    f = QgsFeature()
    f.setAttributes([row[1],row[2]])
    temp.addFeature(f)
    print(row)
# saving changes and adding the layer
temp.commitChanges()
QgsProject.instance().addMapLayer(temp)

The above example can be optimized.