Question

我正在尝试在SQLite数据库中构建一些速度变量。我从下表开始：

CREATE TABLE transactions (
  id TEXT PRIMARY KEY,
  grp TEXT,
  ts NUMERIC
);
BEGIN TRANSACTION;
INSERT INTO exclusions VALUES (
  '1751d4ad-0690-44d7-a13d-1106f48c7761',
  'A',
  STRFTIME ('%s', '2016-01-01 00:00:00')
);
-- ...
END TRANSACTION;
CREATE INDEX grp ON transactions (grp, ts);

然后对于每个id，我想看看在最后x秒内同一grp中发生了多少次CREATE TABLE v_grp ( id TEXT PRIMARY KEY, count_1m NUMERIC, count_7m NUMERIC, count_40m NUMERIC, count_4h NUMERIC, count_1d NUMERIC, count_1w NUMERIC ); BEGIN TRANSACTION; INSERT INTO v_grp SELECT tx1.id AS id, SUM (CASE WHEN tx1.ts - tx2.ts <= 60 THEN 1 ELSE 0 END) AS count_1m, SUM (CASE WHEN tx1.ts - tx2.ts <= 420 THEN 1 ELSE 0 END) AS count_7m, SUM (CASE WHEN tx1.ts - tx2.ts <= 2400 THEN 1 ELSE 0 END) AS count_40m, SUM (CASE WHEN tx1.ts - tx2.ts <= 14400 THEN 1 ELSE 0 END) AS count_4h, SUM (CASE WHEN tx1.ts - tx2.ts <= 86400 THEN 1 ELSE 0 END) AS count_1d, SUM (CASE WHEN tx1.ts - tx2.ts <= 604800 THEN 1 ELSE 0 END) AS count_1w FROM transactions as tx1 JOIN transactions as tx2 ON tx1.grp = tx2.grp AND tx1.ts > tx2.ts AND tx1.ts <= tx2.ts + 604800 GROUP BY tx1.id ; END TRANSACTION;。为此，我做了一个永远运行的自我加入。

$string = 'Andreas';
$reversedString = '';

for($i = mb_strlen($string) - 1; $i >= 0; $i--){
    $reversedString .= $string[$i];
}

我认为这是可以控制的，因为我有超过1年的13M交易，有6k组（最大的一个是交易的10％，第十大是1％）。

我有什么方法可以让这种运行更快，例如，更好的索引或改进选择？

SQL中的Velocity变量

0 个答案: