Hi! I have a collection with

symbols_to_index: []

. I tried dropping and recreating a field to use

symbols_to_index: ["+"]

, but when I search the collection with

q="something+"

, the

+

symbol is not included in the match. Then, I created another collection with

symbols_to_index: ["+"]

, and in this case the

+

symbol was included in the match. Is it possible to set

symbols_to_index

for a single field, or do I need to create a new collection? I am using version 29.0

This has been fixed on the latest HEAD of Typesense, but a release candidate isn't yet available. We'll let you know once we have one

Ok, thank you! Do you have a rough idea of when the release might be available? I’m asking so I can plan whether to create a workaround with an alias or just wait

Not too far, could be as early as MOnday

Hi, @Fanis Tharropoulos! Is there already a release or release candidate available for that fix?

Try out the RC17 candidate!

It seems the fix didn’t work. I tried dropping the field and reindexing twice, but the search is still ignoring the “+” character. Here’s how the field is currently configured:

{
...
    {
      "facet": false,
      "index": true,
      "infix": false,
      "locale": "",
      "name": "title",
      "optional": false,
      "sort": true,
      "stem": false,
      "stem_dictionary": "",
      "store": true,
      "symbols_to_index": [
        "+"
      ],
      "type": "string"
    }
  ],
  "name": "items",
  "num_documents": 1883,
  "symbols_to_index": [],
  "token_separators": []
}

But when I searched, out of 7 results the only one containing the “+” character appears last:

Hi Davi, I'm not sure we support alter + symbols_to_index, you would need to create another collection with it added and then reindex the documents.

Hi Allan, I created a new collection and reindexed all the documents, setting

symbols_to_index

at the field level, using the following schema:

{
  "created_at": 1761228770,
  "default_sorting_field": "title",
  "enable_nested_fields": false,
  "fields": [
    {
      "facet": false,
      "index": true,
      "infix": false,
      "locale": "",
      "name": "title",
      "optional": false,
      "sort": true,
      "stem": false,
      "stem_dictionary": "",
      "store": true,
      "symbols_to_index": ["+"]
      "type": "string"
    },
    ...
  ],
  "name": "drug_items",
  "num_documents": 1902,
  "symbols_to_index": [],
  "token_separators": []
}

However, it still doesn’t work. It only works when I set

symbols_to_index

at the collection level. I’m currently using v30.0.rc25.

Hey Davi, Could you share a few documents that aren't present in the response for the field-level

symbols_to_index

? We'll test it on a local cluster to debug.

Sure. I used the Typesense Cloud interface to search for

"fps 50+"

, and I got seven results. The first one is:

{
  "document": {
    "title": "Helioderm Suncare Facial FPS 50"
  },
  "highlight": {
    "title": {
      "matched_tokens": [
        "FPS",
        "50"
      ],
      "snippet": "Helioderm Suncare Facial <mark>FPS</mark> <mark>50</mark>",
      "value": "Helioderm Suncare Facial <mark>FPS</mark> <mark>50</mark>"
    }
  },
  "highlights": [
    {
      "field": "title",
      "matched_tokens": [
        "FPS",
        "50"
      ],
      "snippet": "Helioderm Suncare Facial <mark>FPS</mark> <mark>50</mark>",
      "value": "Helioderm Suncare Facial <mark>FPS</mark> <mark>50</mark>"
    }
  ],
  "text_match": 1157451471441100800,
  "text_match_info": {
    "best_field_score": "2211897868288",
    "best_field_weight": 15,
    "fields_matched": 1,
    "num_tokens_dropped": 0,
    "score": "1157451471441100921",
    "tokens_matched": 2,
    "typo_prefix_score": 0
  }
}

And the last one is:

{
  "document": {
    "title": "Isdin Fluid Tattoo FPS 50+"
  },
  "highlight": {
    "title": {
      "matched_tokens": [
        "FPS",
        "50"
      ],
      "snippet": "Isdin Fluid Tattoo <mark>FPS</mark> <mark>50</mark>+",
      "value": "Isdin Fluid Tattoo <mark>FPS</mark> <mark>50</mark>+"
    }
  },
  "highlights": [
    {
      "field": "title",
      "matched_tokens": [
        "FPS",
        "50"
      ],
      "snippet": "Isdin Fluid Tattoo <mark>FPS</mark> <mark>50</mark>+",
      "value": "Isdin Fluid Tattoo <mark>FPS</mark> <mark>50</mark>+"
    }
  ],
  "text_match": 1157451437081362400,
  "text_match_info": {
    "best_field_score": "2211881091072",
    "best_field_weight": 15,
    "fields_matched": 1,
    "num_tokens_dropped": 0,
    "score": "1157451437081362553",
    "tokens_matched": 2,
    "typo_prefix_score": 1
  }
}