Updating dateparts of timestamp using query update and/or reindex/pipeline

[petlit2049]

I have log-data from various sources that have been pre-indexed into "master indices". I'd like to re-use that data by copying/re-indexing it into new indices but with parts of the timestamp updated - year, month, day to be specific - we, need the time part to remain equal - only the year/month/day should be updated to the current (re-indexing) date. Can this be accomplished using painless and/or re-indexing pipeline?

[petlit2049]

I've tried (so far) something in the line of this (incomplete). I'm far from a painless guru so please bare with me:

POST _reindex
{
  "source": {
    "index": "packetbeat-*"
  },
  "dest": {
    "index": "<new index>"
  },
  "script": {
    "lang": "painless",
    "source": "def sf = new SimpleDateFormat(\"yyyy-MM-dd'T'HH:mm:ss\"); def dt = sf.parse(ctx._source['@timestamp']); def calendar = sf.getCalendar(); calendar.setTime(dt); def instant = calendar.toInstant(); def localDateTime = LocalDateTime.ofInstant(instant, ZoneOffset.UTC); ctx._source['@timestamp'] = localDateTime.plusDays(...);"
  }
}

The "magic" needs to happen at the end of the painless script. I need all timestamps to be calculated relative to today's date. Assuming today's date is 2024-01-23:

2023-12-09 09:53:12 -> 2024-01-22 09:53:12
2023-12-08 12:53:12 -> 2024-01-23 12:53:12

and so forth.