How to display a Pandas data frame with PyQt5/PySide2

Question

I have a problem with the line below self.tableView.set??????????(df) that supposed to display the data frame in PyQt5. I put ??? there where I am missing the code I need.

def btn_clk(self):
        path = self.lineEdit.text()
        df = pd.read_csv(path)
        self.tableView.set??????????(df)

The rest of the code works, because if I use print(df) in the above code, the data frame is printed in the IPython console. So, Pandas reads the CSV and prints it.

But, I tried many things to get it displayed in PyQt5 and nothing works. I am not very familiar with PyQt, just started to play around with it and I am stuck here.

Here is my code:

from PyQt5 import QtCore, QtGui, QtWidgets
import pandas as pd
class Ui_MainWindow(object):
    def setupUi(self, MainWindow):
        MainWindow.setObjectName("MainWindow")
        MainWindow.resize(662, 512)
        self.centralwidget = QtWidgets.QWidget(MainWindow)
        self.centralwidget.setObjectName("centralwidget")
        self.horizontalLayout = QtWidgets.QHBoxLayout(self.centralwidget)
        self.horizontalLayout.setObjectName("horizontalLayout")
        self.verticalLayout = QtWidgets.QVBoxLayout()
        self.verticalLayout.setObjectName("verticalLayout")
        self.lineEdit = QtWidgets.QLineEdit(self.centralwidget)
        self.lineEdit.setObjectName("lineEdit")
        self.verticalLayout.addWidget(self.lineEdit)
        self.tableView = QtWidgets.QTableView(self.centralwidget)
        self.tableView.setObjectName("tableView")
        self.verticalLayout.addWidget(self.tableView)
        self.pushButton = QtWidgets.QPushButton(self.centralwidget)
        self.pushButton.setObjectName("pushButton")
        self.verticalLayout.addWidget(self.pushButton)
        self.horizontalLayout.addLayout(self.verticalLayout)
        MainWindow.setCentralWidget(self.centralwidget)
        self.menubar = QtWidgets.QMenuBar(MainWindow)
        self.menubar.setGeometry(QtCore.QRect(0, 0, 662, 21))
        self.menubar.setObjectName("menubar")
        MainWindow.setMenuBar(self.menubar)
        self.statusbar = QtWidgets.QStatusBar(MainWindow)
        self.statusbar.setObjectName("statusbar")
        MainWindow.setStatusBar(self.statusbar)

        self.retranslateUi(MainWindow)
        QtCore.QMetaObject.connectSlotsByName(MainWindow)

    def retranslateUi(self, MainWindow):
        _translate = QtCore.QCoreApplication.translate
        MainWindow.setWindowTitle(_translate("MainWindow", "MainWindow"))
        self.pushButton.setText(_translate("MainWindow", "PushButton"))


        self.pushButton.clicked.connect(self.btn_clk)

        MainWindow.show()

    def btn_clk(self):
        path = self.lineEdit.text()
        df = pd.read_csv(path)
        self.tableView.set????????????(df)


if __name__ == "__main__":
    import sys
    app = QtWidgets.QApplication(sys.argv)
    MainWindow = QtWidgets.QMainWindow()
    ui = Ui_MainWindow()
    ui.setupUi(MainWindow)
    MainWindow.show()
    sys.exit(app.exec_())

Answer 1

In the case of QTableView the data must be provided through a model since it implements the MVC (Model-View-Controller) paradigm, in the case of pandas there is no default model but we can create a custom as shown in the following part:

class PandasModel(QtCore.QAbstractTableModel): 
    def __init__(self, df = pd.DataFrame(), parent=None): 
        QtCore.QAbstractTableModel.__init__(self, parent=parent)
        self._df = df.copy()

    def toDataFrame(self):
        return self._df.copy()

    def headerData(self, section, orientation, role=QtCore.Qt.DisplayRole):
        if role != QtCore.Qt.DisplayRole:
            return QtCore.QVariant()

        if orientation == QtCore.Qt.Horizontal:
            try:
                return self._df.columns.tolist()[section]
            except (IndexError, ):
                return QtCore.QVariant()
        elif orientation == QtCore.Qt.Vertical:
            try:
                # return self.df.index.tolist()
                return self._df.index.tolist()[section]
            except (IndexError, ):
                return QtCore.QVariant()

    def data(self, index, role=QtCore.Qt.DisplayRole):
        if role != QtCore.Qt.DisplayRole:
            return QtCore.QVariant()

        if not index.isValid():
            return QtCore.QVariant()

        return QtCore.QVariant(str(self._df.ix[index.row(), index.column()]))

    def setData(self, index, value, role):
        row = self._df.index[index.row()]
        col = self._df.columns[index.column()]
        if hasattr(value, 'toPyObject'):
            # PyQt4 gets a QVariant
            value = value.toPyObject()
        else:
            # PySide gets an unicode
            dtype = self._df[col].dtype
            if dtype != object:
                value = None if value == '' else dtype.type(value)
        self._df.set_value(row, col, value)
        return True

    def rowCount(self, parent=QtCore.QModelIndex()): 
        return len(self._df.index)

    def columnCount(self, parent=QtCore.QModelIndex()): 
        return len(self._df.columns)

    def sort(self, column, order):
        colname = self._df.columns.tolist()[column]
        self.layoutAboutToBeChanged.emit()
        self._df.sort_values(colname, ascending= order == QtCore.Qt.AscendingOrder, inplace=True)
        self._df.reset_index(inplace=True, drop=True)
        self.layoutChanged.emit()

And then use it:

def btn_clk(self):
    path = self.lineEdit.text()
    df = pd.read_csv(path)
    model = PandasModel(df)
    self.tableView.setModel(model)

The complete code is here

Update 03-07-2019:

Some Pandas methods are deprecated so I have implemented a new version (which can also be used in QML as this answer shows):

class DataFrameModel(QtCore.QAbstractTableModel):
    DtypeRole = QtCore.Qt.UserRole + 1000
    ValueRole = QtCore.Qt.UserRole + 1001

    def __init__(self, df=pd.DataFrame(), parent=None):
        super(DataFrameModel, self).__init__(parent)
        self._dataframe = df

    def setDataFrame(self, dataframe):
        self.beginResetModel()
        self._dataframe = dataframe.copy()
        self.endResetModel()

    def dataFrame(self):
        return self._dataframe

    dataFrame = QtCore.pyqtProperty(pd.DataFrame, fget=dataFrame, fset=setDataFrame)

    @QtCore.pyqtSlot(int, QtCore.Qt.Orientation, result=str)
    def headerData(self, section: int, orientation: QtCore.Qt.Orientation, role: int = QtCore.Qt.DisplayRole):
        if role == QtCore.Qt.DisplayRole:
            if orientation == QtCore.Qt.Horizontal:
                return self._dataframe.columns[section]
            else:
                return str(self._dataframe.index[section])
        return QtCore.QVariant()

    def rowCount(self, parent=QtCore.QModelIndex()):
        if parent.isValid():
            return 0
        return len(self._dataframe.index)

    def columnCount(self, parent=QtCore.QModelIndex()):
        if parent.isValid():
            return 0
        return self._dataframe.columns.size

    def data(self, index, role=QtCore.Qt.DisplayRole):
        if not index.isValid() or not (0 <= index.row() < self.rowCount() \
            and 0 <= index.column() < self.columnCount()):
            return QtCore.QVariant()
        row = self._dataframe.index[index.row()]
        col = self._dataframe.columns[index.column()]
        dt = self._dataframe[col].dtype

        val = self._dataframe.iloc[row][col]
        if role == QtCore.Qt.DisplayRole:
            return str(val)
        elif role == DataFrameModel.ValueRole:
            return val
        if role == DataFrameModel.DtypeRole:
            return dt
        return QtCore.QVariant()

    def roleNames(self):
        roles = {
            QtCore.Qt.DisplayRole: b'display',
            DataFrameModel.DtypeRole: b'dtype',
            DataFrameModel.ValueRole: b'value'
        }
        return roles

Answer 2

To fix the sorting...

from natsort import natsorted, index_natsorted, order_by_index
def sort(self, column, order):
    if order == 0:
        self._dataframe = self._dataframe.reindex(index=order_by_index(self._dataframe.index, index_natsorted(self._dataframe[column])))
    else:
        self._dataframe = self._dataframe.reindex(index=order_by_index(self._dataframe.index, reversed(index_natsorted(self._dataframe[column]))))

    self._dataframe.reset_index(inplace=True, drop=True)
    self.setDataFrame(self._dataframe)