結合 sed 與 mv 批量修改文件名

使用 wget 下載數據後發現文件名全帶了鏈接的 query 符號：

$ ls
download?fn=%2FPCAWG%2Fclinical_and_histology%2Fpcawg_donor_clinical_August2016_v9.xlsx
download?fn=%2FPCAWG%2Fclinical_and_histology%2Fpcawg_donor_subtype_cohort_list.xlsx
download?fn=%2FPCAWG%2Fclinical_and_histology%2Fpcawg_specimen_histology_August2016_v9.xlsx
download?fn=%2FPCAWG%2Fconsensus_cnv%2Fconsensus.20170119.somatic.cna.annotated.tar.gz
download?fn=%2FPCAWG%2Fconsensus_cnv%2Fconsensus.20170119.somatic.cna.icgc.public.tar.gz
download?fn=%2FPCAWG%2Fconsensus_cnv%2Fconsensus.20170119.somatic.cna.tcga.public.tar.gz
download?fn=%2FPCAWG%2Fconsensus_cnv%2Fconsensus.20170217.purity.ploidy.txt.gz
download?fn=%2FPCAWG%2Fconsensus_snv_indel%2Ffinal_consensus_passonly.snv_mnv_indel.icgc.public.maf.gz
download?fn=%2FPCAWG%2Fconsensus_sv%2Ffinal_consensus_sv_bedpe_passonly.icgc.public.tgz
download?fn=%2FPCAWG%2Fconsensus_sv%2Ffinal_consensus_sv_bedpe_passonly.tcga.public.tgz

所以最好把前面的內容去掉，sed 可以使用模式匹配進行文本修改，而 mv 可以重命名文件，我們結合兩者試試。首先用單個文件測試修改方式是否正確：

$ echo download?fn=%2FPCAWG%2Fclinical_and_histology%2Fpcawg_specimen_histology_August2016_v9.xlsx | sed -E 's/.*%2(.*)/\1/'
Fpcawg_specimen_histology_August2016_v9.xlsx

然後檢測下目錄下的所有文件都可以這樣處理：

$ ls | sed -E 's/.*%2(.*)/\1/'
Fpcawg_donor_clinical_August2016_v9.xlsx
Fpcawg_donor_subtype_cohort_list.xlsx
Fpcawg_specimen_histology_August2016_v9.xlsx
Fconsensus.20170119.somatic.cna.annotated.tar.gz
Fconsensus.20170119.somatic.cna.icgc.public.tar.gz
Fconsensus.20170119.somatic.cna.tcga.public.tar.gz
Fconsensus.20170217.purity.ploidy.txt.gz
Ffinal_consensus_passonly.snv_mnv_indel.icgc.public.maf.gz
Ffinal_consensus_sv_bedpe_passonly.icgc.public.tgz
Ffinal_consensus_sv_bedpe_passonly.tcga.public.tgz

實際改名字需要使用 for 循環進行迭代：

$ for f in `ls`; do echo `echo $f | sed -E 's/.*%2(.*)/\1/'`; done
Fpcawg_donor_clinical_August2016_v9.xlsx
Fpcawg_donor_subtype_cohort_list.xlsx
Fpcawg_specimen_histology_August2016_v9.xlsx
Fconsensus.20170119.somatic.cna.annotated.tar.gz
Fconsensus.20170119.somatic.cna.icgc.public.tar.gz
Fconsensus.20170119.somatic.cna.tcga.public.tar.gz
Fconsensus.20170217.purity.ploidy.txt.gz
Ffinal_consensus_passonly.snv_mnv_indel.icgc.public.maf.gz
Ffinal_consensus_sv_bedpe_passonly.icgc.public.tgz
Ffinal_consensus_sv_bedpe_passonly.tcga.public.tgz

上面在實際調用 mv 之前檢測一下這樣操作不會有問題，然後修改爲實際要重命名的操作。

$ for f in `ls`; do mv $f `echo $f | sed -E 's/.*%2(.*)/\1/'`; done
$ ls
Fconsensus.20170119.somatic.cna.annotated.tar.gz            Ffinal_consensus_sv_bedpe_passonly.icgc.public.tgz
Fconsensus.20170119.somatic.cna.icgc.public.tar.gz          Ffinal_consensus_sv_bedpe_passonly.tcga.public.tgz
Fconsensus.20170119.somatic.cna.tcga.public.tar.gz          Fpcawg_donor_clinical_August2016_v9.xlsx
Fconsensus.20170217.purity.ploidy.txt.gz                    Fpcawg_donor_subtype_cohort_list.xlsx
Ffinal_consensus_passonly.snv_mnv_indel.icgc.public.maf.gz  Fpcawg_specimen_histology_August2016_v9.xlsx

最後本文可以抽象出來的一個通用操作是：

for f in `ls`; do <cmd> `echo $f | sed -E <operation>; done

該模板可以應用於其他想要進行先修改文件名然後運行命名的操作。