学步园

最近做了一个文本
CDR
转换器，从
A
样式转换为
B
样式，如下所示。

A

样式

RECORD<br /> #input_id 1282892739x001_0011058<br /> #output_id<br /> #input_type GGSNPDPRecord<br /> #output_type GGSNPDPRecord<br /> #addkey<br /> #source_id ggsn01<br /> #filename ggsn01_20100827121321_13251<br /> F recordType 13<br /> F servedIMSI 454063302785690<br /> F ggsniPBinV4Address 203078047018<br /> F chargingID 143CFE18<br /> F sgsniPBinV4Address 203126001225<br /> F accessPointName 11736D6172746F6E652D766F6461666F6E65<br /> F pdpType 0121<br /> F servediPBinV4Address 010052130231<br /> F dynamicAddressFlag FF<br /> F dataVolumeGPRSUplink 76F4<br /> F dataVolumeGPRSDownlink 09B605<br /> F changeCondition 02<br /> F changeTime 1008271213222B0800<br /> F recordOpeningTime 1008271212382B0800<br /> F duration 2C<br /> F causeForRecClosing 00<br /> F recordSequenceNumber 03<br /> F nodeID 6767736E3031<br /> F identifier 04007F00050202000000010001<br /> F changeTime 1008271213222B0800<br /> F serviceIdentifier 54<br /> F sivolumeUplink 711C<br /> F sivolumeDownlink 09B2C5<br /> F serviceIdentifier 62<br /> F sivolumeUplink 05D8<br /> F sivolumeDownlink 0340<br /> F serviceClass 62<br /> F volumeUplink 05D8<br /> F volumeDownlink 0340<br /> F serviceClass 4B<br /> F volumeUplink 711C<br /> F volumeDownlink 09B2C5<br /> F apnSelectionMode 00<br /> F servedMSISDN 198612398426688<br /> F chargingCharacteristics 0800<br /> F chChSelectionMode 00<br /> F sgsnPLMNIdentifier 525F01<br /> F servedIMEI 0121610002016205<br /> F rATType 01<br /> F localSequenceNumber 0DDEF73B<br /> .<br />

B

样式

Record (1) "GGSNPDPRecord"<br /> "recordType" = "13"<br /> "servedIMSI" = "454063302785690"<br /> "ggsniPBinV4Address" = "203078047018"<br /> "chargingID" = "143CFE18"<br /> "sgsniPBinV4Address" = "203126001225"<br /> "accessPointName" = "11736D6172746F6E652D766F6461666F6E65"<br /> "pdpType" = "0121"<br /> "servediPBinV4Address" = "010052130231"<br /> "dynamicAddressFlag" = "FF"<br /> "dataVolumeGPRSUplink" = "76F4"<br /> "dataVolumeGPRSDownlink" = "09B605"<br /> "changeCondition" = "02"<br /> "changeTime" = "1008271213222B0800"<br /> "recordOpeningTime" = "1008271212382B0800"<br /> "duration" = "2C"<br /> "causeForRecClosing" = "00"<br /> "recordSequenceNumber" = "03"<br /> "nodeID" = "6767736E3031"<br /> "identifier" = "04007F00050202000000010001"<br /> "changeTime" = "1008271213222B0800"<br /> "serviceIdentifier" = "54"<br /> "sivolumeUplink" = "711C"<br /> "sivolumeDownlink" = "09B2C5"<br /> "serviceIdentifier" = "62"<br /> "sivolumeUplink" = "05D8"<br /> "sivolumeDownlink" = "0340"<br /> "serviceClass" = "62"<br /> "volumeUplink" = "05D8"<br /> "volumeDownlink" = "0340"<br /> "serviceClass" = "4B"<br /> "volumeUplink" = "711C"<br /> "volumeDownlink" = "09B2C5"<br /> "apnSelectionMode" = "00"<br /> "servedMSISDN" = "198612398426688"<br /> "chargingCharacteristics" = "0800"<br /> "chChSelectionMode" = "00"<br /> "sgsnPLMNIdentifier" = "525F01"<br /> "servedIMEI" = "0121610002016205"<br /> "rATType" = "01"<br /> "localSequenceNumber" = "0DDEF73B"<br /> End of Record (1)<br />

A

样式到

B

样式的转换要点如下：

1.    


header
和
trailer
都可直接转换，但是
B
样式需要加上
CDR
的计数

2.    


B
样式的
header
还含有
record type
，需要从
A
样式的某个注释行中获取

3.    


字段域的转换比较简单，都是一行对一行

4.    


A
样式有注释行，转换时要全部删除

转换由

awk

和

sed

完成，完整的命令是这样的：

awk 'BEGIN { cnt=0; } { if ( $0 == "RECORD" ) { cnt++;
printf("Record (%d)/n", cnt); } else if ( $0 == "." ) {
printf("End of Record (%d)/n", cnt); } else { print $0; } }' layout_a
| sed '/^Record/{N;N;N;s//n//g;}' | sed 's/^/(Record (.*)/).*input_type
/(.*/)$//1 "/2"/;s/^F /(.*/) /(.*/)$/  
"/1" = "/2"/;/^#/d' > layout b

具体解析如下：

1.   

awk 'BEGIN { cnt=0; } {
if ( $0 == "RECORD" ) { cnt++; printf("Record (%d)/n",
cnt); } else if ( $0 == "." ) { printf("End of Record
(%d)/n", cnt); } else { print $0; } }'

layout_a | sed
'/^Record/{N;N;N;s//n//g;}' | sed 's/^/(Record (.*)/).*input_type /(.*/)$//1
"/2"/;s/^F /(.*/) /(.*/)$/  

"/1" = "/2"/;/^#/d' > layout b

判断
CDR
的
header
和
trailer
。如果是
header
，即该行内容为
”RECORD”
，计数器加一，且输出
”Record (?)”
到下一步；如果是
trailer
，即该行内容为
”.”
，输出
”End of Record (?)”
到下一步；否则，不作修改输出到下一步。

2.   


awk 'BEGIN { cnt=0; } { if ( $0 == "RECORD" ) { cnt++;
printf("Record (%d)/n", cnt); } else if ( $0 == "." ) {
printf("End of Record (%d)/n", cnt); } else { print $0; } }' layout_a
|
sed
'/^Record/{N;N;N;s//n//g;}'

| sed 's/^/(Record (.*)/).*input_type /(.*/)$//1
"/2"/;s/^F /(.*/) /(.*/)$/  

"/1" = "/2"/;/^#/d' > layout b

遇到
header
，合并后续三行。因为第四行保留了该
CDR
的
Record type
。

3.   


awk 'BEGIN { cnt=0; } { if ( $0 == "RECORD" ) { cnt++;
printf("Record (%d)/n", cnt); } else if ( $0 == "." ) {
printf("End of Record (%d)/n", cnt); } else { print $0; } }' layout_a
| sed '/^Record/{N;N;N;s//n//g;}' | sed '
s/^/(Record (.*)/).*input_type
/(.*/)$//1 "/2"/

;
s/^F /(.*/) /(.*/)$/  

"/1" = "/2"/;/^#/d' > layout b

从上一步生成的
header
提取合适的内容，形成
B
样式的
header
。注意括号的运用。

4.   


awk 'BEGIN { cnt=0; } { if ( $0 == "RECORD" ) { cnt++;
printf("Record (%d)/n", cnt); } else if ( $0 == "." ) {
printf("End of Record (%d)/n", cnt); } else { print $0; } }' layout_a
| sed '/^Record/{N;N;N;s//n//g;}' | sed 's/^/(Record (.*)/).*input_type
/(.*/)$//1 "/2"/;
s/^F /(.*/) /(.*/)$/  

"/1" = "/2"/;

/^#/d' > layout b

提取各个字段的键和值，并加上双引号，形成
B
样式的字段域。同样注意括号的运用。

5.   


awk 'BEGIN { cnt=0; } { if ( $0 == "RECORD" ) { cnt++;
printf("Record (%d)/n", cnt); } else if ( $0 == "." ) {
printf("End of Record (%d)/n", cnt); } else { print $0; } }' layout_a
| sed '/^Record/{N;N;N;s//n//g;}' | sed 's/^/(Record (.*)/).*input_type
/(.*/)$//1 "/2"/;s/^F /(.*/) /(.*/)$/  
"/1" = "/2"/;
/^#/d

' > layout b

删除剩余的注释行。