Question

我正在尝试过滤 X 列并获取所有缺失的周数据。

例如，一个 df（样本日期，在实际 df 中我们将拥有整周的数据）：

mean_date         column X  
2021-04-01        x_123
2021-04-01        y_324
2021-04-02        x_123
2021-04-03        x_123

我需要找到本周对应的 X 列的所有缺失日期，即

result_df：

mean_date_missing     column_X
2021-03-28            x_123
2021-03-29            x_123  
2021-03-30            x_123
2021-03-31            x_123
..
2021-03-28            y_324
2021-03-29            y_324  
2021-03-30            y_324
2021-03-31            y_324
2021-04-02            y_324
2021-04-03            y_324

Answer 1

可能不是最优雅的方法，但我认为这可行：

首先，我会得到该周所有日期的列表，我们称之为 #include <iostream> #include <thread> #include <vector> #include <atomic> #include <cstdlib> #define _WINSOCK_DEPRECATED_NO_WARNINGS #include <winsock2.h> #include <Windows.h> #include <ws2tcpip.h> #pragma comment(lib, "Ws2_32.lib") std::atomic<bool> stopRunning = false; void server() { int iResult; sockaddr_in listen_address; listen_address.sin_family = AF_INET; listen_address.sin_port = htons(1000); listen_address.sin_addr.S_un.S_addr = INADDR_ANY; SOCKET listen_socket = socket(AF_INET, SOCK_STREAM, 0); if (listen_socket == INVALID_SOCKET) { iResult = WSAGetLastError(); std::cerr << "socket failed: " << iResult << "\n"; return; } u_long mode = 1; if (ioctlsocket(listen_socket, FIONBIO, &mode) == SOCKET_ERROR) { iResult = WSAGetLastError(); std::cerr << "ioctlsocket failed: " << iResult << "\n"; closesocket(listen_socket); return; } if (bind(listen_socket, (sockaddr*)&listen_address, sizeof(listen_address)) == SOCKET_ERROR) { iResult = WSAGetLastError(); std::cerr << "bind failed: " << iResult << "\n"; closesocket(listen_socket); return; } if (listen(listen_socket, SOMAXCONN) == SOCKET_ERROR) { iResult = WSAGetLastError(); std::cerr << "listen failed: " << iResult << "\n"; closesocket(listen_socket); return; } std::vector<SOCKET> clients; SOCKET client_socket; sockaddr client_info; int addrlen; fd_set rfds; timeval timeout; char buffer[1024]; while (!stopRunning.load()) { FD_ZERO(&rfds); FD_SET(listen_socket, &rfds); for(auto sckt : clients) { FD_SET(sckt, &rfds); } timeout.tv_sec = 1; timeout.tv_usec = 0; iResult = select(0, &rfds, NULL, NULL, &timeout); if (iResult < 0) { iResult = WSAGetLastError(); std::cerr << "select failed: " << iResult << "\n"; break; } if (iResult == 0) continue; if (FD_ISSET(listen_socket, &rfds)) { addrlen = sizeof(client_info); client_socket = accept(listen_socket, (struct sockaddr*)&client_info, &addrlen); if (client_socket == INVALID_SOCKET) { iResult = WSAGetLastError(); std::cerr << "accept failed: " << iResult << "\n"; break; } if (ioctlsocket(client_socket, FIONBIO, &mode) == SOCKET_ERROR) { iResult = WSAGetLastError(); std::cerr << "ioctlsocket failed: " << iResult << "\n"; closesocket(client_socket); } else { std::cout << "New connection" << std::endl; clients.push_back(client_socket); } } for (size_t i = 0; i < clients.size();) { client_socket = clients[i]; if (FD_ISSET(client_socket, &rfds)) { iResult = recv(client_socket, buffer, sizeof(buffer), 0); if (iResult == SOCKET_ERROR) { iResult = WSAGetLastError(); std::cerr << "recv failed: " << iResult << "\n"; closesocket(client_socket); clients.erase(clients.begin()+i); continue; } if (iResult == 0) { std::cout << "Closing connection" << std::endl; closesocket(client_socket); clients.erase(clients.begin()+i); continue; } std::cout.write(buffer, iResult); std::cout << std::endl; } ++i; } } //clean up closesocket(listen_socket); for (auto sckt : clients) { closesocket(sckt); } } int main() { WSADATA data; int iResult = WSAStartup(MAKEWORD(2, 2), &data); if (iResult != 0) { std::cerr << "WSAStartup failed: " << iResult << "\n"; return 0; } std::thread server_thread(server); std::system("pause"); stopRunning = true; server_thread.join(); WSACleanup(); return 0; }。您可以手动执行此操作，也可以为其编写一个函数（如果您使用谷歌搜索，您应该找到解决方案）。所以，你现在有这样的东西：

<data-in name="test_feed" dataset="input_data">
     <start-instance>${coord:current(-(coord:hoursInDay(0))}</start-instance>
     <end-instance>${coord:current(0)}</end-instance>
</data-in>

接下来，我将按 weekdays 对 DataFrame 进行分组，因为您想分别为该列的每个可能值找到缺失的天数。

weekdays = ["2021-04-01", "2021-04-02", "2021-04-03", "2021-04-04", "2021-04-05"]

然后，迭代它以找到每组缺失的日期：

column X

现在，所有必要的信息都存储在 grouped = df.groupby(df["column x"]) 中。您只需要从中制作一个 DataFrame：

missing_list = []
for key, item in grouped:
    existing_dates = item["mean_date"].to_list()
    missing_dates = np.setdiff1d(weekdays, existing_dates)
    for date in missing_dates:
        missing_list.append([date, key])

熊猫根据列的缺失日期过滤行

1 个答案: