实现MySQL表数据全量索引和增量索引,基于Solr DIH组件实现起来比较简单,只需要重复使用Solr的DIH(Data Import Handler)组件,对data-config.xml进行简单的修改即可。Solr DIH组件的实现类为org.apache.solr.handler.dataimport.DataImportHandler,在Solr的solrconfig.xml中配置两个handler,配置分别说明如下。
全量索引
solrconfig.xml配置如下:
1 |
< requestHandler name = "/dataimport" |
2 |
class = "org.apache.solr.handler.dataimport.DataImportHandler" > |
4 |
< str name = "config" >data-config.xml</ str > |
上面这个是针对全量索引的,主要是配置data-config.xml文件,示例如下所示:
02 |
< dataSource name = "jdbc" driver = "com.mysql.jdbc.Driver" |
04 |
user = "developer" password = "sedept@shiyanjun.cn" /> |
05 |
< document name = "mkt_data" > |
06 |
< entity name = "marketing_data" pk = "id" |
07 |
query = "select * from marketing_data limit ${dataimporter.request.length} offset ${dataimporter.request.offset}" |
08 |
transformer = "RegexTransformer" > |
09 |
< field column = "id" name = "id" /> |
10 |
< field column = "domain" name = "domain" /> |
11 |
< field column = "alex_rank" name = "alex_rank" /> |
12 |
< field column = "server_port" name = "server_port" /> |
13 |
< field column = "cert_validity_notBefore" name = "cert_validity_notBefore" /> |
14 |
< field column = "cert_validity_notAfter" /> |
15 |
< field column = "cert_validity_notAfter_yyyyMMdd" regex = "(.*?)\s+.*" name = "cert_validity_notAfter_yyyyMMdd" sourceColName = "cert_validity_notAfter" /> |
16 |
< field column = "cert_issuer_brand" name = "cert_issuer_brand" /> |
17 |
< field column = "cert_validation" name = "cert_validation" /> |
18 |
< field column = "cert_isMultiDomain" name = "cert_isMultiDomain" /> |
19 |
< field column = "cert_issuer_brand_isXRelated" name = "cert_issuer_brand_isXRelated" /> |
20 |
< field column = "cert_isWildcard" name = "cert_isWildcard" /> |
21 |
< field column = "cert_notAfter" name = "cert_notAfter" /> |
22 |
< field column = "special_ssl" name = "special_ssl" /> |
23 |
< field column = "competitor_logo" name = "competitor_logo" /> |
24 |
< field column = "segment" name = "segment" /> |
上面主要是通过内置变量 “${dataimporter.request.length}”和 “${dataimporter.request.offset}”来设置一个批次索引的数据表记录数,请求的URL示例如下:
上面表示,对数据表中id范围为[10000000, 1100000]的记录进行索引,因为数据表可能达到千万记录数,而且线上有业务在操作数据库,所以要选择分批进行索引。
增量索引
solrconfig.xml配置如下:
01 |
< requestHandler name = "/deltaimport" class = "org.apache.solr.handler.dataimport.DataImportHandler" > |
03 |
< str name = "config" >delta-data-config.xml</ str > |
06 |
上面定义请求的接口为“deltaimport”,对应的配置文件delta- data-config.xml内容如下所示: |
08 |
< dataSource name = "jdbc" driver = "com.mysql.jdbc.Driver" |
10 |
user = "developer" password = "sedept@shiyanjun.cn" /> |
11 |
< document name = "mkt_data" > |
12 |
< entity name = "marketing_data" pk = "id" |
13 |
query = "select * from marketing_data" |
14 |
deltaImportQuery = "select * from marketing_data where id='${dih.delta.id}'" |
15 |
deltaQuery = "select id from marketing_data where updated_at > '${dih.last_index_time}'" |
16 |
transformer = "RegexTransformer" > |
17 |
< field column = "id" name = "id" /> |
18 |
< field column = "domain" name = "domain" /> |
19 |
< field column = "alex_rank" name = "alex_rank" /> |
20 |
< field column = "server_port" name = "server_port" /> |
21 |
< field column = "cert_validity_notBefore" name = "cert_validity_notBefore" /> |
22 |
< field column = "cert_validity_notAfter" /> |
23 |
< field column = "cert_validity_notAfter_yyyyMMdd" regex = "(.*?)\s+.*" name = "cert_validity_notAfter_yyyyMMdd" sourceColName = "cert_validity_notAfter" /> |
24 |
< field column = "cert_issuer_brand" name = "cert_issuer_brand" /> |
25 |
< field column = "cert_validation" name = "cert_validation" /> |
26 |
< field column = "cert_isMultiDomain" name = "cert_isMultiDomain" /> |
27 |
< field column = "cert_issuer_brand_isXRelated" name = "cert_issuer_brand_isXRelated" /> |
28 |
< field column = "cert_isWildcard" name = "cert_isWildcard" /> |
29 |
< field column = "cert_notAfter" name = "cert_notAfter" /> |
30 |
< field column = "special_ssl" name = "special_ssl" /> |
31 |
< field column = "competitor_logo" name = "competitor_logo" /> |
32 |
< field column = "segment" name = "segment" /> |
上面主要是通过内置变量“${dih.delta.id}”和 “${dih.last_index_time}”来记录本次索引的id和最后索引时间。这里,会保存在deltaimport.properties文件中,示例如下:
1 |
#Sat Feb 16 17:48:49 CST 2013 |
2 |
last_index_time=2013-02-16 17\:48\:48 |
3 |
marketing_data.last_index_time=2013-02-16 17\:48\:48 |
请求增量索引的接口为“deltaimport”,可以写一个定时脚本,例如每隔一天执行一次 (一天索引一次,只索引数据表中更新的记录),请求URL示例如下:
1 |
172.0.8.212:8080/search-server/core0/dataimport?command=delta-import |
有关“query”,“deltaImportQuery”, “deltaQuery”的解释,引用官网说明,如下所示:
- The query gives the data needed to populate fields of the Solr document in full-import
- The deltaImportQuery gives the data needed to populate fields when running a delta-import
- The deltaQuery gives the primary keys of the current entity which have changes since the last index time
还有更灵活的索引方式,以后再讲。