如何在不影响性能的情况下获取 App Engine 数据存储区查询中每个实体的游标?
How to get the cursor of each entity in a App Engine datastore query without performance hit?
我有一个使用游标 (Objectify v5) 的数据存储区查询,我想在结果列表中的每个项目之后获取游标。代码如下所示:
public List<Puzzle> queryWithCursor(String cursor, String order, int limit) {
Query<Puzzle> query = ObjectifyService.ofy()
.load()
.type(Puzzle.class)
.order(order)
.limit(limit);
query = query.startAt(Cursor.fromWebSafeString(cursor));
List<Puzzle> puzzles = new ArrayList<>();
QueryResultIterator<Puzzle> iterator = query.iterator();
while (iterator.hasNext()) {
Puzzle puzzle = iterator.next();
puzzle.setCursor(iterator.getCursor().toWebSafeString());
puzzles.add(puzzle);
}
return puzzles;
}
虽然该方法工作正常,但它在幕后触发了如此多的数据存储区查询。基本上,每次 iterator.getCursor() 运行时,它都会触发一个额外的查询。我从 Stackdriver Trace 了解到,如果 limit 为 20,该方法总共会触发 19 个查询(似乎最后一个 .getCursor() 不会触发额外的查询)。因此,这种方法比使用偏移量的类似查询更慢且成本更高。
这真的是一个错误吗?有没有办法避免性能下降?
import com.google.appengine.api.datastore.Cursor;
import com.google.appengine.api.datastore.DatastoreService;
import com.google.appengine.api.datastore.DatastoreServiceFactory;
import com.google.appengine.api.datastore.Entity;
import com.google.appengine.api.datastore.FetchOptions;
import com.google.appengine.api.datastore.PreparedQuery;
import com.google.appengine.api.datastore.Query;
import com.google.appengine.api.datastore.Query.SortDirection;
import com.google.appengine.api.datastore.QueryResultList;
import java.io.IOException;
import java.io.PrintWriter;
import javax.servlet.ServletException;
import javax.servlet.http.HttpServlet;
import javax.servlet.http.HttpServletRequest;
import javax.servlet.http.HttpServletResponse;
public class ListPeopleServlet extends HttpServlet {
static final int PAGE_SIZE = 15;
private final DatastoreService datastore;
public ListPeopleServlet() {
datastore = DatastoreServiceFactory.getDatastoreService();
}
@Override
protected void doGet(HttpServletRequest req, HttpServletResponse resp)
throws ServletException, IOException {
FetchOptions fetchOptions = FetchOptions.Builder.withLimit(PAGE_SIZE);
// If this servlet is passed a cursor parameter, let's use it.
String startCursor = req.getParameter("cursor");
if (startCursor != null) {
fetchOptions.startCursor(Cursor.fromWebSafeString(startCursor));
}
Query q = new Query("Person").addSort("name", SortDirection.ASCENDING);
PreparedQuery pq = datastore.prepare(q);
QueryResultList<Entity> results;
try {
results = pq.asQueryResultList(fetchOptions);
} catch (IllegalArgumentException e) {
// IllegalArgumentException happens when an invalid cursor is used.
// A user could have manually entered a bad cursor in the URL or there
// may have been an internal implementation detail change in App Engine.
// Redirect to the page without the cursor parameter to show something
// rather than an error.
resp.sendRedirect("/people");
return;
}
resp.setContentType("text/html");
resp.setCharacterEncoding("UTF-8");
PrintWriter w = resp.getWriter();
w.println("<!DOCTYPE html>");
w.println("<meta charset=\"utf-8\">");
w.println("<title>Cloud Datastore Cursor Sample</title>");
w.println("<ul>");
for (Entity entity : results) {
w.println("<li>" + entity.getProperty("name") + "</li>");
}
w.println("</ul>");
String cursorString = results.getCursor().toWebSafeString();
// This servlet lives at '/people'.
w.println("<a href='/people?cursor=" + cursorString + "'>Next page</a>");
}
}
这实际上是数据存储的基本行为,至少在旧的 sdk 中是这样(与 Objectify 6 使用的新 sdk 不同,后者可能相同也可能不同)。在非批边界调用 getCursor() 会重新启动查询。你可以用低级的API.
试试
有一个解决方法:制作您自己的 Cursor class。它应该由低级 Cursor 和一个偏移量组成。显式设置 chunk()
大小,然后您的游标应由索引 0 处的 Cursor
加上块中的偏移量组成。
然后当您想在该游标处重新启动查询时,请使用 .cursor(batchStartCursor).offset(offsetIntoBatch)
。
我有一个使用游标 (Objectify v5) 的数据存储区查询,我想在结果列表中的每个项目之后获取游标。代码如下所示:
public List<Puzzle> queryWithCursor(String cursor, String order, int limit) {
Query<Puzzle> query = ObjectifyService.ofy()
.load()
.type(Puzzle.class)
.order(order)
.limit(limit);
query = query.startAt(Cursor.fromWebSafeString(cursor));
List<Puzzle> puzzles = new ArrayList<>();
QueryResultIterator<Puzzle> iterator = query.iterator();
while (iterator.hasNext()) {
Puzzle puzzle = iterator.next();
puzzle.setCursor(iterator.getCursor().toWebSafeString());
puzzles.add(puzzle);
}
return puzzles;
}
虽然该方法工作正常,但它在幕后触发了如此多的数据存储区查询。基本上,每次 iterator.getCursor() 运行时,它都会触发一个额外的查询。我从 Stackdriver Trace 了解到,如果 limit 为 20,该方法总共会触发 19 个查询(似乎最后一个 .getCursor() 不会触发额外的查询)。因此,这种方法比使用偏移量的类似查询更慢且成本更高。
这真的是一个错误吗?有没有办法避免性能下降?
import com.google.appengine.api.datastore.Cursor;
import com.google.appengine.api.datastore.DatastoreService;
import com.google.appengine.api.datastore.DatastoreServiceFactory;
import com.google.appengine.api.datastore.Entity;
import com.google.appengine.api.datastore.FetchOptions;
import com.google.appengine.api.datastore.PreparedQuery;
import com.google.appengine.api.datastore.Query;
import com.google.appengine.api.datastore.Query.SortDirection;
import com.google.appengine.api.datastore.QueryResultList;
import java.io.IOException;
import java.io.PrintWriter;
import javax.servlet.ServletException;
import javax.servlet.http.HttpServlet;
import javax.servlet.http.HttpServletRequest;
import javax.servlet.http.HttpServletResponse;
public class ListPeopleServlet extends HttpServlet {
static final int PAGE_SIZE = 15;
private final DatastoreService datastore;
public ListPeopleServlet() {
datastore = DatastoreServiceFactory.getDatastoreService();
}
@Override
protected void doGet(HttpServletRequest req, HttpServletResponse resp)
throws ServletException, IOException {
FetchOptions fetchOptions = FetchOptions.Builder.withLimit(PAGE_SIZE);
// If this servlet is passed a cursor parameter, let's use it.
String startCursor = req.getParameter("cursor");
if (startCursor != null) {
fetchOptions.startCursor(Cursor.fromWebSafeString(startCursor));
}
Query q = new Query("Person").addSort("name", SortDirection.ASCENDING);
PreparedQuery pq = datastore.prepare(q);
QueryResultList<Entity> results;
try {
results = pq.asQueryResultList(fetchOptions);
} catch (IllegalArgumentException e) {
// IllegalArgumentException happens when an invalid cursor is used.
// A user could have manually entered a bad cursor in the URL or there
// may have been an internal implementation detail change in App Engine.
// Redirect to the page without the cursor parameter to show something
// rather than an error.
resp.sendRedirect("/people");
return;
}
resp.setContentType("text/html");
resp.setCharacterEncoding("UTF-8");
PrintWriter w = resp.getWriter();
w.println("<!DOCTYPE html>");
w.println("<meta charset=\"utf-8\">");
w.println("<title>Cloud Datastore Cursor Sample</title>");
w.println("<ul>");
for (Entity entity : results) {
w.println("<li>" + entity.getProperty("name") + "</li>");
}
w.println("</ul>");
String cursorString = results.getCursor().toWebSafeString();
// This servlet lives at '/people'.
w.println("<a href='/people?cursor=" + cursorString + "'>Next page</a>");
}
}
这实际上是数据存储的基本行为,至少在旧的 sdk 中是这样(与 Objectify 6 使用的新 sdk 不同,后者可能相同也可能不同)。在非批边界调用 getCursor() 会重新启动查询。你可以用低级的API.
试试有一个解决方法:制作您自己的 Cursor class。它应该由低级 Cursor 和一个偏移量组成。显式设置 chunk()
大小,然后您的游标应由索引 0 处的 Cursor
加上块中的偏移量组成。
然后当您想在该游标处重新启动查询时,请使用 .cursor(batchStartCursor).offset(offsetIntoBatch)
。