当我们索引的mapping,setting发生变更时,我们需要重建索引来使得这些变更生效。es提供了两种方式来完成重建索引的操作:
1:update by query,在本索引重建
2:reindex,在新索引上重建
我们通过具体实例来分别看下。
核心点,如果不执行update_by_query,修改mapping前将会无法查询到。执行update_by_query的方式如下:
POST {索引名称}/_update_by_query
{
}
具体实例(从上往下看即可)
:
# 1:创建索引,指定mapping
DELETE blogs
PUT blogs
{
"mappings": {
"properties": {
"content": {
"type": "text"
},
"keyword": {
"type": "text"
}
}
}
}
# 2:修改mapping前插入一条数据
PUT blogs/_doc/1
{
"content": "Hadoop is coll",
"keyword": "Hadoop"
}
# 3:修改mapping,增加子字段english
PUT blogs/_mapping
{
"properties": {
"content": {
"type": "text",
"fields": {
"english": {
"type": "text",
"analyzer": "english"
}
}
},
"keyword": {
"type": "text"
}
}
}
# 4:修改mapping后插入一条记录
PUT blogs/_doc/2
{
"content": "Elasticsearch is hard to study,but i will keep up",
"keyword": "Elasticsearch"
}
# 5:查询修改mapping前的数据,查询不到
POST blogs/_search
{
"query": {
"match": {
"content.english": "Hadoop"
}
}
}
# 6:查询修改mapping后的数据,可以查询到
POST blogs/_search
{
"query": {
"match": {
"content.english": "Elasticsearch"
}
}
}
# 7:执行update_by_query,重建本索引
POST blogs/_update_by_query
{
}
# 8:再次查询修改mapping前的数据,就可以查询到了
POST blogs/_search
{
"query": {
"match": {
"content.english": "Hadoop"
}
}
}
es不允许修改字段类型,想要完成修改字段类型操作的话必须使用reindex。
# 1:创建索引,指定mapping,注意此时keyword字段是text类型
DELETE blogs
PUT blogs
{
"mappings": {
"properties": {
"content": {
"type": "text"
},
"keyword": {
"type": "text"
}
}
}
}
# 2:插入几条测试数据
POST blogs/_bulk
{"index": {"_id": 1}}
{"content":"Hadoop is coll","keyword":"Hadoop"}
{"index": {"_id": 2}}
{"content":"Elasticsearch is hard to study,but i will keep up","keyword":"Elasticsearch"}
# 3:尝试修改keyword的数据类型从text为keyword,会报错
# mapper [keyword] of different type, current_type [text], merged_type [keyword]
PUT blogs/_mapping
{
"properties": {
"content": {
"type": "text"
},
"keyword": {
"type": "keyword"
}
}
}
# 4:创建全新的索引,但是指定keyword字段类型为keyword
DELETE blogs_fix
PUT blogs_fix
{
"mappings": {
"properties": {
"content": {
"type": "text"
},
"keyword": {
"type": "keyword"
}
}
}
}
# 5:通过reindex拷贝数据到新索引
POST _reindex
{
"source": {
"index": "blogs"
},
"dest": {
"index": "blogs_fix"
}
}
# 6:看是否成功 "count" : 2即为成功了
GET blogs_fix/_count
# 7:对keyword字段执行term agg,可以成功,进一步验证其字段类型是keyword(因为textfielddata默认是关闭的无法term)
GET blogs_fix/_search
{
"aggs": {
"term分组验证就是keyword类型": {
"terms": {
"field": "keyword"
}
}
}
}
第二点需要注意,提前设置好mapping和setting,要不然可能会做无用功。
如果是目标index已经存在部分文档,为了忽略已存在的文档,可增加该参数:
通过增加wait_for_completion=false
,使用异步方式来复制索引,可通过GET _tasks?detailed=true&actions=*reindex
来查看任务执行情况: