mysql分区表 查询分区的数据行数与实际不符

[xiaoyan]

mysql按月分区：
PARTITION s201612 VALUES LESS THAN (TO_DAYS('2017-01-01 ')),
PARTITION s201701 VALUES LESS THAN (TO_DAYS('2017-02-01')),
PARTITION s201702 VALUES LESS THAN (TO_DAYS('2017-03-01 ')),
PARTITION s201703 VALUES LESS THAN (TO_DAYS('2017-04-01 ')),
PARTITION s201704 VALUES LESS THAN (TO_DAYS('2017-05-01 ')),
PARTITION s201705 VALUES LESS THAN (TO_DAYS('2017-06-01 ')),
PARTITION s201706 VALUES LESS THAN (TO_DAYS('2017-07-01')),
PARTITION s201707 VALUES LESS THAN (TO_DAYS('2017-08-01')),
PARTITION s201708 VALUES LESS THAN (TO_DAYS('2017-09-01')),
PARTITION s201709 VALUES LESS THAN (TO_DAYS('2017-10-01')),
PARTITION s201710 VALUES LESS THAN (TO_DAYS('2017-11-01')),
PARTITION s201711 VALUES LESS THAN (TO_DAYS('2017-12-01')),
PARTITION s201712 VALUES LESS THAN (TO_DAYS('2018-01-01')),
PARTITION p2 VALUES LESS THAN MAXVALUE

通过 INFORMATION_SCHEMA.partitions 表查询分区的数据行数，查到s201701行数65条，
SELECT
partition_name part,
partition_expression expr,
partition_description descr,
table_rows
FROM
INFORMATION_SCHEMA.partitions
WHERE
TABLE_SCHEMA = schema()
AND TABLE_NAME='table_ym';

实际上一月份的数据有77条，删除s2017这个分区，发现一月份的数据已经全部删除。如何让查询表分区的数据条数和实际的数据条数一致？ps：发现12条差距刚好是2017年1月1日这天的数据。

[xiaoyan]

查看mysql官方文档，发现 INFORMATION_SCHEMA.partitions的TABLE_ROWS字段其实提供的就是一个大概数值。附上官方文档说明 TABLE_ROWS: The number of table rows in the partition. For partitioned InnoDB tables, the row count given in the TABLE_ROWS column is only an estimated value used in SQL optimization, and may not always be exact.