为提高效率，提问时请提供以下信息，问题描述清晰可优先响应。
【DM版本】:DM9
【操作系统】：CENTOS7
【CPU】: INTEL
【问题描述】*：
如下sql执行效率差, 在oracle中执行很快

 SELECT
        m.s_company_code AS companyCode,
        d.n_period_code AS period,
        m.s_contents_code AS contentsCode,
        sum(NVL(d.N_TOTAL_VALUE_OPEN, 0)) AS totalValueOpen,
        sum(NVL(d.n_total_value_close, 0)) AS totalValueClose
FROM
        AM_MASTER_DATA m
LEFT JOIN AM_DEPRECIATION D ON
        m.S_COMPANY_CODE = d.S_COMPANY_CODE
        AND d.N_PERIOD_CODE IN ( 202301,202302 )
        AND d.S_ASSETS_CODE = m.S_ASSETS_CODE
WHERE
        m.s_company_code IN ( '2BJ1101030' , '2BJ1102030' , '2BJ1506070' )
        AND ( m.S_CONTENTS_CODE LIKE '010301' || '%' )
        AND d.S_ID IS NOT NULL
GROUP BY
        m.s_company_code,
        d.n_period_code,
        m.s_contents_code

查看执行计划如下:

1   #NSET2: [1, 1, 378] 
2     #PRJT2: [1, 1, 378]; exp_num(5), is_atom(FALSE) 
3       #HAGR2: [1, 1, 378]; grp_num(3), sfun_num(2), distinct_flag[0,0]; slave_empty(0) keys(DMTEMPVIEW_889372821.TMPCOL0, DMTEMPVIEW_889372821.TMPCOL1, DMTEMPVIEW_889372821.TMPCOL2) 
4         #PRJT2: [1, 22, 378]; exp_num(5), is_atom(FALSE) 
5           #SLCT2: [1, 22, 378]; (D.N_PERIOD_CODE IN LIST AND D.S_COMPANY_CODE IN LIST)
6             #HASH2 INNER JOIN: [1, 22, 378];  KEY_NUM(2); KEY(M.S_COMPANY_CODE=D.S_COMPANY_CODE AND M.S_ASSETS_CODE=D.S_ASSETS_CODE) KEY_NULL_EQU(0, 0)
7               #SLCT2: [1, 22, 378]; D.S_ASSETS_CODE = M.S_ASSETS_CODE
8                 #NEST LOOP INDEX JOIN2: [1, 22, 378] 
9                   #ACTRL: [1, 22, 378];
10                    #NEST LOOP INDEX JOIN2: [1, 11, 144] 
11                      #CONST VALUE LIST: [1, 3, 48]; row_num(3), col_num(1)
12                      #BLKUP2: [1, 3, 144]; AM_MASTER_DATA_S_COMPANY_CODE_IDX(M)
13                        #SSEK2: [1, 3, 144]; scan_type(ASC), AM_MASTER_DATA_S_COMPANY_CODE_IDX(AM_MASTER_DATA as M), scan_range[(DMTEMPVIEW_889372823.colname,'010301'),(DMTEMPVIEW_889372823.colname,'010302'))
14                  #BLKUP2: [1, 2, 126]; INDEX33627091(D)
15                    #SLCT2: [1, 2, 126]; (D.N_PERIOD_CODE IN LIST AND D.S_COMPANY_CODE IN LIST)
16                      #SSEK2: [1, 2, 126]; scan_type(ASC), INDEX33627091(AM_DEPRECIATION as D), scan_range[(M.S_COMPANY_CODE,min,min),(M.S_COMPANY_CODE,max,max))
17              #CSCN2: [778, 5095517, 234]; INDEX33627003(AM_DEPRECIATION as D); btr_scan(1)

我发现第17行描述的计划中, 对AM_DEPRECIATION表做了全表扫描,且成本很高, 不知道为啥要在left join后又进行了高成本的hash join,
请问如何优化?

我还发现inner join的执行效率都很差. 有什么系统参数需要调整吗? 请给出建议!,万分感谢