shell里怎么用正则表达式取子串

hty_1984

比如从以下字符串中取出<mediaUrl type="string">rtmpte://rhapsodyev-159.fcod.llnwd.net/a4376/v1/s/9/5/1/9/1/384419159?e=1345174409&amp;h=ee604ab75c6577196ae4d3dd1ef27b6c</mediaUrl>中的rtmpte://rhapsodyev-159.fcod.llnwd.net/a4376/v1/s/9/5/1/9/1/384419159?e=1345174409&amp，有什么好方法，grep,sed,awk都行，越简单越好！


<?xml version="1.0" encoding="UTF-8"?>
<o><data class="object"><accountType type="string">RHAPSODY_TOGO</accountType><bitRate type="number">128</bitRate><format type="string">mp3 64</format><freePlaysRemaining type="number">0</freePlaysRemaining><genre class="object"><genreId type="string">g.115</genreId><name type="string">Pop</name><rootGenreName type="string"/></genre><mediaUrl type="string">rtmpte://rhapsodyev-159.fcod.llnwd.net/a4376/v1/s/9/5/1/9/1/384419159?e=1345174409&amp;h=ee604ab75c6577196ae4d3dd1ef27b6c</mediaUrl><playbackSessionId type="string">1345173808914:jason-huo@rhapsody.com:40134:rn_popout</playbackSessionId><playbackType type="string">NORMAL</playbackType><purchaseInfo class="object"><type type="string">track</type><url type="string">http://mp3.rhapsody.com/goto?rcid=Tra.30867607&amp;pcode=rn&amp;addToCart=Tra.30867607</url></purchaseInfo><radVersion type="number">0</radVersion><track class="object"><album class="object"><albumArt162X162Url type="string">http://static.rhap.com/img/170x170/3/1/6/9/2279613_170x170.jpg</albumArt162X162Url><albumArt70X70Url type="string"/><albumId type="string">Alb.30848999</albumId><albumType type="number">0</albumType><compilation type="boolean">false</compilation><copyright type="string"/><displayName type="string">The Fame Monster</displayName><displayableAlbumTypes class="array"><e type="string">Edited</e></displayableAlbumTypes><label type="string"/><name type="string">The Fame Monster</name><numberOfDiscs type="string">1</numberOfDiscs><parentalAuthority type="boolean">false</parentalAuthority><primaryArtistDisplayName type="string">Lady GaGa</primaryArtistDisplayName><primaryArtistId type="string">Art.20067373</primaryArtistId><primaryStyle type="string">Dance Pop</primaryStyle><releaseYear type="string">2009</releaseYear><rightFlags type="number">33594</rightFlags><shortcut type="string">lady-gaga/the-fame-monster</shortcut><tracks class="array"/></album><albumId type="string">Alb.30848999</albumId><artist class="object"><artistId type="string">Art.20067373</artistId><genre class="object" null="true"/><imageURL type="string"/><name type="string">Lady Gaga</name><rightFlags type="number">33594</rightFlags><shortcut type="string">lady-gaga</shortcut></artist><artistId type="string">Art.20067373</artistId><blurbs class="array"/><discIndex type="string">1</discIndex><displayAlbumName type="string">The Fame Monster</displayAlbumName><displayArtistName type="string">Lady Gaga</displayArtistName><genre class="object"><genreId type="string">g.115</genreId><name type="string">Pop</name><rootGenreName type="string"/></genre><genreId type="string">g.290</genreId><name type="string">Poker Face</name><playbackSeconds type="number">237</playbackSeconds><price type="string"/><purchaseInfo class="object"><type type="string">track</type><url type="string">http://mp3.rhapsody.com/goto?rcid=Tra.30867607&amp;pcode=rn&amp;addToCart=Tra.30867607</url></purchaseInfo><rightFlags type="number">33338</rightFlags><sampleUrl type="string">http://listen.vo.llnwd.net/g1/5/1/2/9/1/384419215.mp3</sampleUrl><trackId type="string">Tra.30867607</trackId><trackIndex type="string">12</trackIndex></track><unlimited type="boolean">true</unlimited><userInfo class="object" null="true"/><validTerritory type="boolean">true</validTerritory></data><status class="object"><blocked type="boolean">false</blocked><errorCode type="string"/><errorMessage type="string"/><mid type="string">g1bS=fAxjQCKKudM8GX8dsjzzBQVWnVU</mid><success type="boolean">true</success></status></o>

liion631818

1. 
   
2. sed -n 's/.*<mediaUrl[^>]*>\([^<;]*\)[^<]*<\/mediaUrl>.*/\1/gp'
   

复制代码

hty_1984

还真是可以了，能具体解释一下什么意思吗？我是shell新手，多谢了！！！
回复 2# liion631818


   

Shell_HAT

回复 3# hty_1984


可以先把教程读一遍
http://bbs.chinaunix.net/thread-452942-1-1.html

Ray001

文本处理，还是perl更强。我更喜欢用perl。

rdcwayx

1. grep -oP '(?<=<mediaUrl type="string">)[^<;]+' infile

复制代码