Timeline for Mysql full text search with other indexed data
Current License: CC BY-SA 4.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Mar 5, 2020 at 16:53 | comment | added | Rick James | @Echotest - You should construct the query based on what options the use selects. In particular, we have discussed the query with EventName is present; the query without it should look very different (no FT, no LIKE "%..."). I'll study the indexes further later. (I'll be out for a while.) Probably it will involve a few indexes starting with event_log_domain_id, then having identifier/Customer/Type. But I am not sure. How selective is each? | |
| Mar 5, 2020 at 16:06 | comment | added | Echotest | was my update helpful? should I use different set of indexes? | |
| Mar 2, 2020 at 17:22 | comment | added | Echotest | I grabbed old staging schema instead of the latest prod schema. It's updated correctly now. My apologies | |
| Feb 27, 2020 at 15:09 | comment | added | Echotest | also added more context on how we actually query the data | |
| Feb 26, 2020 at 23:41 | comment | added | Echotest | I updated question with create table statement. I don't have any text columns in that table, but there is anywhere between 11-20 million records living in that table for 45 days and they slowly die (TTL expires) of and are populated with another set of events that are there for 45 days. Sometimes there are more or less, but it's a big table, and we do filtering on top for troubleshooting purposes. | |
| Feb 25, 2020 at 6:16 | history | answered | Rick James | CC BY-SA 4.0 |