Question

我们假设我们在DB中有以下表格：

CREATE TABLE #nodes(reporter varchar(10), groupname varchar(10), node varchar(20))
CREATE TABLE #nodes_extract(id int IDENTITY, min_node varchar(20), count_nodes int, descr varchar(10) );
CREATE TABLE #nodes_histo(reporter varchar(10), groupname varchar(10), node varchar(20), nodes_extract_id int)

数据示例：

INSERT INTO #nodes (reporter, groupname, node) VALUES ('rep1','group1','node1')
INSERT INTO #nodes (reporter, groupname, node) VALUES ('rep1','group1','node2')
INSERT INTO #nodes (reporter, groupname, node) VALUES ('rep1','group2','node3')
INSERT INTO #nodes (reporter, groupname, node) VALUES ('rep2','group1','node1')
INSERT INTO #nodes (reporter, groupname, node) VALUES ('rep2','group1','node4')
INSERT INTO #nodes (reporter, groupname, node) VALUES ('rep2','group2','node5')
INSERT INTO #nodes (reporter, groupname, node) VALUES ('rep3','group1','node5')
INSERT INTO #nodes (reporter, groupname, node) VALUES ('rep3','group1','node6')
INSERT INTO #nodes (reporter, groupname, node) VALUES ('rep3','group2','node7')

让我们将一些数据提取到#nodes_extract

中

INSERT INTO #nodes_extract
SELECT min(node), count(node), 'blabla'
FROM #nodes
GROUP BY reporter, groupname

现在我想将#nodes中的所有信息插入#nodes_histo并添加nodes_extract_id #nodes_histo中的预期结果：

    reporter    groupname   node    nodes_extract_id
    rep1        group1      node1   1
    rep1        group1      node2   1
    rep1        group2      node3   4
    rep2        group1      node1   2
    rep2        group1      node4   2
    rep2        group2      node5   5
    rep3        group1      node5   3
    rep3        group1      node6   3
    rep3        group2      node7   6

你将如何实现目标？是否可以不改变表格结构？

非常感谢！ KAFE

Answer 1

嗯，我觉得我找到了办法..

首先，在insert命令中需要DISTINCT：

INSERT INTO #nodes_extract
SELECT DISTINCT min(node), count(node), 'blabla'
FROM #nodes
GROUP BY reporter, groupname

通过以下选择，我可以提取我正在寻找的信息：

INSERT INTO #nodes_histo
SELECT a.reporter, a.groupname,a.node, e.id 
FROM 
    (SELECT  reporter
            , groupname
            , node
            , min(node) OVER (PARTITION BY REPORTER, groupname) m
            , count(node) OVER (PARTITION BY REPORTER, groupname) c
    FROM #nodes) a
LEFT JOIN 
    #nodes_extract e
ON a.m = e.min_node
AND a.c = e.count_nodes

Answer 2

不，您可能需要包含一种加入表的方法，可能是在#nodes_extract中包含reporter和groupname。即使您尝试使用相同的查询（以及其他列）重新生成#nodes_extract，也无法保证订单是相同的。

数据库表提取和链接提取的数据

2 个答案: