iamlaosong文
将文本文件导入到数据库,如果不需要对文件进行处理,用sqlldr导入不仅方便,而且速度快。下面这个导入脚本,通过循环生成每个文本文件导入的控制文件,其中X'09'表示分隔符是TAB符,然后再调用sqlldr命令进行导入。
#import datafile that unzip by timing job: jsunzip.sh #eidt by iamlaosong 20170926 #usage: ./jsimp.sh 20170101 BCF # $1=OPDATE $2=SYF|BCF BAKDIR=/home/oracle/ems_suan CTLFILE=control.ctl for tp in CK HK JK LY TD ZY do tbname=EMSAPP_JS_${tp}_$2_TEST txtfile="${BAKDIR}/$1/$1_340000_1_${tp}_$2.txt" # process txtfile echo "${txtfile} --> ${tbname}" if [ -f ${txtfile} ]; then # generate control file echo load data>${CTLFILE} echo infile "'${txtfile}'">>${CTLFILE} echo append into table $tbname>>${CTLFILE} echo fields terminated by "X'09'">>${CTLFILE} # echo optionally enclosed by "'\"'">>${CTLFILE} if [ X"${tp}" == X"CK" ] || [ X"${tp}" == X"JK" ]; then echo "(JSRQ Date 'YYYYMMDD',ZBTM,YJTM,YJZLBH,YJZLMC,ZBZL,ZBZLMC,JDS,JSSX,ZLSX,JIFEIZL,JIHEZL,CLJSFDM,CLJSF,CLJDSDM,CLJDS,CLJXS,CLJ,CLJGS,SJJSFDM,SJJSF,SJJDSDM,SJJDS,SJJXSDM,SJJXS,SJJDM,SJJ,SJJGS,YCXX,FEILI,FEIYONG,MARK)">>${CTLFILE} elif [ X"${tp}" == X"HK" ] || [ X"${tp}" == X"LY" ]; then echo "(JSRQ Date 'YYYYMMDD',YLDM,YLMC,LDLSH,YLJB,YLZL,FCZDM,FCZ,JSZDM,JSZ,YJJDM,YJJ,JDJDM,JDJ,HBBH,LICHENG,PYJSFDM,PYJSF,PYJDSDM,PYJDS,PYJXSDM,PYJXS,PYJDM,PYJ,PYJGS,MARK)">>${CTLFILE} elif [ X"${tp}" == X"TD" ]; then echo "(JSRQ Date 'YYYYMMDD',PYDQ,SFTT,TDJSFDM,TDJSF,TDJDSDM,TDJDS,TDJXSDM,TDJXS,TDJDM,TDJ,TDJGS,SZFEILI,XZFEILI,MARK)">>${CTLFILE} elif [ X"${tp}" == X"ZY" ]; then echo "(JSRQ Date 'YYYYMMDD',CLJXSDM,CLJDM,MARK)">>${CTLFILE} fi # import data sqlldr userid=emssxjk/emssxjk control=control.ctl skip=1 cat control.log >> impsuan.log # read -p "txt file: press enter to continue..." var fi done
shell脚本中逻辑或的判断除了代码中的写法外,还可以采用-o的方法,即:
if [ X"${tp}" == X"CK" -o X"${tp}" == X"JK" ]; then
判断字符串中加上一个X是防止变量tp为空时出现语法错误,本例中tp是循环变量,不会出现空值,但这种写法比较好,所以保留着。
原文链接:https://www.f2er.com/oracle/207575.html