admin管理员组

文章数量:1287527

I have a Node.js application that interacts with a PostgreSQL database containing millions of users. For a specific feature, I need to fetch around 100,000 users based on a criterion (for example, users that have a particular channel id in their tracking configuration) and then process each user (e.g., creating notification or autotrade tasks).

Currently, my implementation fetches all matching users into memory and then processes them in chunks, like so:

async function processMessageForSubscribers(channelId, channelName, message, addresses) {
  try {
  
    // get around 100000 users in memory
    const users = await getUsersByTrackedTelegramChannel(channelId);
    const CHUNK_SIZE = 500;
    const notifyTasks = [];
    const autotradeTasks = [];

    // Split users into chunks for parallel processing
    const processUserChunk = async (userChunk) => {
      await Promise.all(
        userChunk.map(async (user) => {
          const config = user.trackingConfig[channelId];
          const autotradeAmount = config?.autotradeAmount;
          
          if (config.newPost === 'NOTIFY') {
            // Create notification tasks
            createNotificationTask(user, addresses, message, channelId, channelName, autotradeAmount, notifyTasks);
          }
          
        })
      );
    };

    // Process users in chunks
    for (let i = 0; i < users.length; i += CHUNK_SIZE) {
      const chunk = users.slice(i, i + CHUNK_SIZE);
      await processUserChunk(chunk);
    }

    await queueTasks(notifyTasks, autotradeTasks);
  } catch (error) {
    console.error('Error processing subscribers:', error);
    throw error;
  }
}

本文标签：

版权声明：本文标题：javascript - How can I efficiently process large PostgreSQL datasets in Node.js without high memory overhead? - Stack Overflow 内容由网友自发贡献，该文观点仅代表作者本人， 转载请联系作者并注明出处：http://www.betaflare.com/web/1741313960a2371807.html， 本站仅提供信息存储空间服务，不拥有所有权，不承担相关法律责任。如发现本站有涉嫌抄袭侵权/违法违规的内容，一经查实，本站将立刻删除。