linux - 在 unix 中处理日期格式

Question

下面的sql代码将被插入到一个文本文件中，作为input.txt的输入:

insert into t_agg_sgsn_gprs_usage
select
    served_imsi as imsi,
    served_msisdn as msisdn,
    NVL(cell_identity,'-1') as cell_id,
    NVL(trim(to_char(location_area_code,'99999999')),'UNKNOWN'),
    substr(served_imei,1,8) tac_code,
    to_char(to_timestamp(record_opening_time,'YYYYMMDDHH24MISS'),'YYYYMMDD')
        as call_dt,
    'ajay' as loaded_dt,
    count(1) as cdr_count,
    sum(data_volume_uplink) as uploaded_data_volume,
    sum(data_volume_downlink) as downloaded_data_volume,
    case
        when substr(imsi,6,2)='65'
        then substr(imsi,1,9)
        else substr(imsi,1,5)
    end imsi_prefix
from
    t_cdrc_sgsn_pdp_18
    where loaded_dt LIKE 'ajay%' AND trim(served_imsi) != ''
        AND trim(served_msisdn) != '' AND trim(served_imei) != ''
    group by imsi, msisdn, tac_code, cell_id, location_area_code, call_dt;

我正在编写一个 unix 脚本，它应该给出日期递增 1 的输出，并附加到一个 sql 文件(changer.sql)，直到指定的日期。

insert into t_agg_sgsn_gprs_usage
select
        served_imsi as imsi,
        served_msisdn as msisdn,
        NVL(cell_identity,'-1') as cell_id,
        NVL(trim(to_char(location_area_code,'99999999')),'UNKNOWN'),
        substr(served_imei,1,8) tac_code,
        to_char(to_timestamp(record_opening_time,'YYYYMMDDHH24MISS'),'YYYYMMDD') as call_dt,
        '2012-10-17' as loaded_dt,
        count(1) as cdr_count,
        sum(data_volume_uplink) as uploaded_data_volume,
        sum(data_volume_downlink) as downloaded_data_volume,
        case when substr(imsi,6,2)='65' then substr(imsi,1,9) else substr(imsi,1,5) end imsi_prefix
from
        t_cdrc_sgsn_pdp_18
        where loaded_dt LIKE '2012-10-17%' AND trim(served_imsi) != '' AND trim(served_msisdn) != '' AND trim(served_imei) != ''
        group by imsi, msisdn, tac_code, cell_id, location_area_code, call_dt;

insert into t_agg_sgsn_gprs_usage
select
        served_imsi as imsi,
        served_msisdn as msisdn,
        NVL(cell_identity,'-1') as cell_id,
        NVL(trim(to_char(location_area_code,'99999999')),'UNKNOWN'),
        substr(served_imei,1,8) tac_code,
        to_char(to_timestamp(record_opening_time,'YYYYMMDDHH24MISS'),'YYYYMMDD') as call_dt,
        '2012-10-18' as loaded_dt,
        count(1) as cdr_count,
        sum(data_volume_uplink) as uploaded_data_volume,
        sum(data_volume_downlink) as downloaded_data_volume,
        case when substr(imsi,6,2)='65' then substr(imsi,1,9) else substr(imsi,1,5) end imsi_prefix
from
        t_cdrc_sgsn_pdp_18
        where loaded_dt LIKE '2012-10-18%' AND trim(served_imsi) != '' AND trim(served_msisdn) != '' AND trim(served_imei) != ''
        group by imsi, msisdn, tac_code, cell_id, location_area_code, call_dt;

并继续递增并追加到 changer.sql 文件中，直到指定的结束日期。

以下是我目前所拥有的，但仍然没有产生预期的结果:

startdate = `date +"%y-%m-%d"`
set startdate =  2012-10-17
enddate = `date “+%y-%m-%d` 
set enddate = 2012-12-22
for($startdate <= $enddate)
do
for file in $(grep -il "ajay" input.txt)
do
sed -e "s/ajay/$startdate/ig" $file >> changer.sql
done
$startdate --startdate="-1 days ago"
done

Answer 1

考虑到你的问题只涉及 bash，你有以下错误:

1 - 你会做什么:

startdate=...
set startdate=...

2 - 你在 bash 中赋值的语法是错误的；之间不应有空格:

rvalue=lvalue
startdate=`date +"%y-%m-%d"`

3 - 此语法对于 for 循环是错误的；我猜你想要一个 while 循环:

for ($startdate <= $enddate)  *wrong*
do ...
done

while (( $startdate <= $enddate )); do
    ...
done

4 - 你对什么有期望:

$startdate --startdate="-1 days ago"

您可能试图将 $startdate 的值减少 1 天，因为您必须这样做:

startdate=$(date -d $startdate' -1 days' +"%y-%m-%d")