R2-37-第二阶段第二次作业

洪学志 2018-02-11 00:23:15 阅读: 1106

一、

library(DBI)

library(RMySQL)

con <- dbConnect(MySQL(),host="localhost",dbname="rdb",user="root",password="")

dbSendQuery(con,'SET NAMES utf8') 

dbGetInfo(con)  

dbListTables(con)  

dbRemoveTable(con,"test") 看老大的来练习

 二、

library(DBI)

library(RMySQL)

con <- dbConnect(MySQL(),host="localhost",dbname="rdb",user="root",password="")

dbSendQuery(con,'SET NAMES utf8')

library(httr)

totalNum=562

pageSize=10

totalPage=ceiling(totalNum/pageSize)

currentPage=1

term='(cell[TA]) AND 2017[DP]'

usehistory='Y'

querykey=''

webenv=''

postSearchUrl='https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esearch.fcgi'

while(currentPage<=totalPage){

  retstart=(currentPage-1)*pageSize

  r <- POST(postSearchUrl, 

            body = list(

              db='pubmed',

              term=term,

              retmode='json',

              retstart=retstart,

              retmax=pageSize,

              usehistory=usehistory,

              rettype='uilist'

            )

      )

  stop_for_status(r)

  data=content(r, "parsed", "application/json")

  esearchresult=data$esearchresult

  querykey=esearchresult$querykey

  webenv=esearchresult$webenv

  idlist =esearchresult$idlist

  n = length(idlist)

  pmid=c()

  i = 1

  while(i<=n){

    pmid=c(pmid, as.character(idlist[i][1]))

    i = i+1

  }

  article=data.frame('pmid'=pmid)

  dbWriteTable(con,"article",article,append=TRUE)

  currentPage = currentPage + 1

}

dbDisconnect(con)

2-1.png

三、

library(RMySQL)

library(xml2)

library(httr)

library(DBI)

con <- dbConnect(MySQL(),host="localhost",dbname="rdb",user="root",password="")

dbSendQuery(con,'SET NAMES utf8')

rs <- dbSendQuery(con, "SELECT * FROM article WHERE isdone=0")

while (!dbHasCompleted(rs)) {

  chunk <- dbFetch(rs, 10) 

  pmidStr=""

  i=1

  n=nrow(chunk)

  while(i<=n){

    pmidStr=paste(pmidStr,chunk[i,3],sep=",")

    i=i+1

  }

    pmidStr=substr(pmidStr,2,100000)

    postFetchUrl='https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi'

  r2 <- POST(postFetchUrl, 

             body = list(

               db='pubmed',

               id=pmidStr,

               retmode='xml'

             )

  )

  stop_for_status(r2)

  data2=content(r2, "parsed", "application/xml")

  article=xml_children(data2)

  count=length(article)

  cnt=1

  while(cnt<=count){

    title=xml_text(xml_find_first(article[cnt],".//ArticleTitle"))

    abstract=xml_text(xml_find_first(article[cnt],".//AbstractText"))

    pmid=xml_text(xml_find_first(article[cnt],".//PMID"))

    title=gsub("'","",title)

    abstract=gsub("'","",abstract)

    sql=paste("UPDATE article SET title='",title,"',abstract='",abstract,"',isdone=1"," where pmid='",pmid,"'",sep="")

    con2 <- dbConnect(MySQL(),host="localhost",dbname="rdb",user="root",password="")

    dbSendQuery(con2,'SET NAMES utf8')

    dbSendQuery(con2,sql)

    dbDisconnect(con2)

    cnt = cnt + 1

    Sys.sleep(1)

      }

}

3-1.png

3-2.png

 

 
邀请讨论

附件

{{f.title}} 大小 {{f.file_size}} 下载 {{f.count_download}} 金币 {{f.count_gold}}
{{item.nick_name}} 受邀请回答 {{item.create_time}}
{{item.refer_comment.nick_name}} {{item.refer_comment.create_time}}

附件

{{f.title}} 大小 {{f.file_size}} 下载 {{f.count_download}} 金币 {{f.count_gold}}
切换到完整回复 发送回复
赞({{item.count_zan}}) 踩({{item.count_cai}}) 删除 回复 关闭
科研狗©2015-2024 科研好助手,京ICP备20005780号-1 建议意见

服务热线

178 0020 3020

微信服务号