ES Presentation Demo Resources

Create Index and Mapping

PUT /es-product-demo
{
  "settings": {
    "number_of_shards": 2,
    "number_of_replicas": 2
  },
  "mappings": {
    "properties": {
      "id": { "type": "keyword" },
      "name": { "type": "text" },
      "tags": { "type": "keyword" },
      "price": { "type": "float" },
      "active": { "type": "boolean" },
      "createdDate": { "type": "date" },
      "variantStock": {
        "type": "nested",
        "properties": {
          "size": { "type": "keyword" },
          "stock": { "type": "integer" }
        }
      },
      "reviews": {
        "properties": {
          "userName": { "type": "text" },
          "rating": { "type": "integer" }
        }
      }
    }
  }
}

Bulk Insert Data

POST /es-product-demo/_bulk
{"index":{"_id":1}}
{"id":1,"name":"Cool Red Shirt","tags":["Shirt","Red"],"price":25.99,"active":true,"createdDate":"2021-09-08","variantStock":[{"size":"Small","stock":15},{"size":"Medium","stock":42},{"size":"Large","stock":3}],"reviews":[{"userName":"John Doe","rating":4},{"userName":"Sarah Leg","rating":3}]}
{"index":{"_id":2}}
{"id":2,"name":"Amazing Jazzy Shorts","tags":["Shorts"],"price":55.50,"active":true,"createdDate":"2021-09-06","variantStock":[{"size":"Small","stock":0},{"size":"Medium","stock":15},{"size":"Large","stock":32}],"reviews":[{"userName":"Bob Builder","rating":5},{"userName":"Sofine Signer","rating":4}]}
{"index":{"_id":3}}
{"id":3,"name":"Red Amazing Hat","tags":["Red","Hat"],"price":29.99,"active":true,"createdDate":"2021-09-06","variantStock":[{"size":"Small","stock":20},{"size":"Medium","stock":1},{"size":"Large","stock":12}],"reviews":[{"userName":"Jonny Jackhammer","rating":2},{"userName":"Bobby Bad","rating":1}]}
{"index":{"_id":4}}
{"id":4,"name":"Supreme Blue Jacket","tags":["Blue","Jacket"],"price":82,"active":false,"createdDate":"2021-09-02","variantStock":[{"size":"Small","stock":0},{"size":"Medium","stock":0},{"size":"Large","stock":3}]}
{"index":{"_id":5}}
{"id":5,"name":"Sneaky Sneakers","tags":["Sneakers"],"price":149.99,"active":true,"createdDate":"2021-09-08","variantStock":[{"size":"Small","stock":35},{"size":"Medium","stock":18},{"size":"Large","stock":2}],"reviews":[{"userName":"Roger Rabbit","rating":2},{"userName":"Mary Maine","rating":5}]}
{"index":{"_id":6}}
{"id":6,"name":"Soft Gloves Red","tags":["Gloves", "Red"],"price":34.80,"active":true,"createdDate":"2021-09-01","variantStock":[{"size":"Small","stock":84},{"size":"Medium","stock":2},{"size":"Large","stock":41}],"reviews":[{"userName":"Supper Man","rating":4},{"userName":"John Doe","rating":2}]}

Get all products

GET /es-product-demo/_search
{
  "query": {
    "match_all": {}
  }
}

Single Match Query

Analyses the search term SHIRT !red! --> ["shirt", "red"], and find all documents that contains 1 or more of these terms in the name field

GET /es-product-demo/_search
{
  "_source": [
    "name",
    "tags",
    "price"
  ],
  "query": {
    "bool": {
      "must": [
        {
          "match": {
            "name": "SHIRT !red!"
          }
        }
      ]
    }
  }
}

Analyse Text Field

Analyses the text SHIRT !red! --> ["shirt", "red"] since the name property is of type text which executes in the Query Context

GET /es-product-demo/_analyze
{
  "field": "name",
  "text": "SHIRT !red!"
}

Analyse HTML with Email Text Field

The goal is to get the text to only contain the upper case version of the emails

Standard Analysing

GET /es-product-demo/_analyze
{
  "text": "<a href='mailto:info@mycompany.com'>Info@mycompany.com | Sales@MyCompany.com</a>",
  "char_filter": [],
  "tokenizer": "standard",
  "filter": ["lowercase"]
}

Strip HTML elements

GET /es-product-demo/_analyze
{
  "text": "<a href='mailto:info@mycompany.com'>Info@mycompany.com | Sales@MyCompany.com</a>",
  "char_filter": ["html_strip"],
  "tokenizer": "standard",
  "filter": ["lowercase"]
}

Use the Email tokenizer

GET /es-product-demo/_analyze
{
  "text": "<a href='mailto:info@mycompany.com'>Info@mycompany.com | Sales@MyCompany.com</a>",
  "char_filter": ["html_strip"],
  "tokenizer": "uax_url_email",
  "filter": ["lowercase"]
}

Uppercase all tokens with a token filter

GET /es-product-demo/_analyze
{
  "text": "<a href='mailto:info@mycompany.com'>Info@mycompany.com | Sales@MyCompany.com</a>",
  "char_filter": ["html_strip"],
  "tokenizer": "uax_url_email", 
  "filter": ["uppercase"]
}

Analyse Keyword Field

Analyses the text SHIRT !red! --> ["SHIRT !red!"] since the tags property is of type keyword which executes in the Filter Context

GET /es-product-demo/_analyze
{
  "field": "tags",
  "text": "SHIRT !red!"
}

Match with Should query to boost relevance scoring

Matches all the documents containing the term red in the name property, and boosts the relevance score if the term soft appears in the name field

GET /es-product-demo/_search
{
  "_source": [
    "name",
    "tags",
    "price"
  ],
  "query": {
    "bool": {
      "must": [
        {
          "match": {
            "name": "red"
          }
        }
      ],
      "should": [
        {
          "match": {
            "name": "SOFT"
          }
        }
      ]
    }
  }
}

Exclude all products with Shirt tag

Same as above, but excludes all documents that contain the term Shirt as a tag

GET /es-product-demo/_search
{
  "_source": [
    "name",
    "tags",
    "price"
  ],
  "query": {
    "bool": {
      "must": [
        {
          "match": {
            "name": "red"
          }
        }
      ],
      "should": [
        {
          "match": {
            "name": "SOFT"
          }
        }
      ],
      "must_not": [
        {
          "term": {
            "tags": {
              "value": "Shirt"
            }
          }
        }
      ]
    }
  }
}

Only show products with price less than or equal to 30

Same as above, but filters documents where the price is less than or equal to 30

GET /es-product-demo/_search
{
  "_source": [
    "name",
    "tags",
    "price"
  ],
  "query": {
    "bool": {
      "must": [
        {
          "match": {
            "name": "red"
          }
        }
      ],
      "should": [
        {
          "match": {
            "name": "SOFT"
          }
        }
      ],
      "must_not": [
        {
          "term": {
            "tags": {
              "value": "Shirt"
            }
          }
        }
      ],
      "filter": [
        {
          "range": {
            "price": {
              "lte": 30
            }
          }
        }
      ]
    }
  }
}

Search Reviews for John with rating 4 or higher

Attempts to find all documents where there is a review made by john with a rating of more than 3. This will not return the expected result as the array is of a normal object type, which does not maintain relationships of the object properties. Thus all documents containing a review by john, and with any review with a rating of 3 or more, will be matched

GET /es-product-demo/_search
{
  "_source": [
    "name",
    "reviews"
  ],
  "query": {
    "bool": {
      "must": [
        {
          "match": {
            "reviews.userName": "JOHN"
          }
        },
        {
          "range": {
            "reviews.rating": {
              "gt": 3
            }
          }
        }
      ]
    }
  }
}

Search Variant Stock for Large with more than 30 in stock

Matches all documents that has the more than 30 Large items

GET /es-product-demo/_search
{
  "_source": [
    "name",
    "variantStock"
  ],
  "query": {
    "bool": {
      "must": [
        {
          "nested": {
            "path": "variantStock",
            "query": {
              "bool": {
                "must": [
                  {
                    "match": {
                      "variantStock.size": "Large"
                    }
                  }
                ],
                "filter": [
                  {
                    "range": {
                      "variantStock.stock": {
                        "gt": 30
                      }
                    }
                  }
                ]
              }
            }
          }
        }
      ]
    }
  }
}

BONUS

Create Index with nested reviews

PUT /es-product-nested-reviews-demo
{
  "settings": {
    "number_of_shards": 2,
    "number_of_replicas": 2
  },
  "mappings": {
    "properties": {
      "id": { "type": "keyword" },
      "name": { "type": "text" },
      "tags": { "type": "keyword" },
      "price": { "type": "float" },
      "active": { "type": "boolean" },
      "createdDate": { "type": "date" },
      "totalStock": { "type": "integer" },
      "averageRating": { "type": "float" },
      "variantStock": {
        "type": "nested",
        "properties": {
          "size": { "type": "keyword" },
          "stock": { "type": "integer" }
        }
      },
      "reviews": {
        "type": "nested",
        "properties": {
          "userName": { "type": "text" },
          "rating": { "type": "integer" }
        }
      }
    }
  }
}

Re-Index using the ReIndex API

Added bonus to calculate newly added fields totalStock and averageRating

POST /_reindex
{
  "source": {
    "index": "es-product-demo"
  },
  "dest": {
    "index": "es-product-nested-reviews-demo"
  },
  "script": {
    "source": """
    def totalStock = 0;
    for (def i = 0; i < ctx._source.variantStock.length; i++) {
      totalStock += ctx._source.variantStock[i].stock;
    }
    ctx._source.totalStock = totalStock;
    
    def totalRatings = 0;
    if (ctx._source.reviews == null) {
      return;
    }
    for (def i = 0; i < ctx._source.reviews.length; i++) {
      totalRatings += ctx._source.reviews[i].rating;
    }
    ctx._source.averageRating = Math.round(totalRatings / ctx._source.reviews.length * 100) / 100;
    """
  }
}

Get all re-indexed documents

GET /es-product-nested-reviews-demo/_search
{
  "query": {
    "match_all": {}
  }
}

Search Reviews for John with rating 4 or higher

This will now return expected results, since reviews are of type nested

GET /es-product-nested-reviews-demo/_search
{
  "_source": [
    "name",
    "reviews"
  ],
  "query": {
    "bool": {
      "must": [
        {
          "nested": {
            "path": "reviews",
            "query": {
              "bool": {
                "must": [
                  {
                    "match": {
                      "reviews.userName": "JOHN"
                    }
                  },
                  {
                    "range": {
                      "reviews.rating": {
                        "gt": 3
                      }
                    }
                  }
                ]
              }
            }
          }
        }
      ]
    }
  }
}