I've a database that stores data read from different sensors. The table looks like this:
[SensorID][timestampMS][value] [Sensor1][123420][10] [Sensor1][123424][15] [Sensor1][123428][6554] [Sensor1][123429][20] What I would like to do is the following: There are some reads that are corrupted (numbers that are 6554), and I would like to Update that with the next value that is not corrupted (in the example shown below that would be 20). So, if a number开发者_开发技巧 is 6554, I would like to update that with the next value (in timestamp), that is not corrupted. I was thinking on doing this in PHP, but I wonder if it's possible to do it directly with a SQL script.Appreciate :)
You can use a correlated sub-query...
UPDATE
myTable
SET
value = (SELECT value FROM myTable AS NextValue WHERE sensorID = myTable.SensorID AND timestampMS > myTable.timestampMS ORDER BY timestampMS ASC LIMIT 1)
WHERE
value = 6554
The sub-query gets all the following results, ordered by timestampMS and takes just the first one; That being the next value for that SensorID.
Note: If no "next" value exists, it will attempt to update with a value of NULL. To get around this, you can add this to the WHERE clause...
AND EXISTS (SELECT value FROM myTable AS NextValue WHERE sensorID = myTable.SensorID AND timestampMS > myTable.timestampMS ORDER BY timestampMS ASC LIMIT 1)
EDIT
Or, to be shorter, just use IFNULL(<sub_query>, value)
...
Not sure if this is valid syntax, can't test it ATM. You may need to change this to be JOIN
s instead of the nested subqueries, but in concept you can do something like (for SQL Server):
UPDATE t1
SET Value = ( SELECT Value
from MyTable t2
WHERE t2.SensorID =t1.SensorID
AND t2.[timestamp] =
( SELECT MIN([TimeStamp])
FROM mytable t3
where t3.sensorid = t2.sensorID
AND t3.[timestamp] > t2.[timestamp]
)
)
FROM Mytable t1
WHERE t1.value = 6554
I did a workaround based on Dems solution, and it works in Mysql:
I've created a "copy" of the sensors table like this:
drop table if exists sensors_new;
create table if not exists sensors_new like sensors;
insert into sensors_new select * from sensors;
Then I do what Dems recommended me doing, but using this new aux table in the select (to avoid the error that Mysql launches when Updating a table while doing a select in the same table).
UPDATE
sensors
SET
raw_data = (SELECT raw_data FROM sensors_new AS NextValue WHERE sensor_id = sensors.sensor_id AND timestampMS > sensors.timestampMS ORDER BY timestampMS ASC LIMIT 1)
WHERE
value = 6554
Then, just drop this auxiliar table.
I hope this helps Mysql users.
精彩评论