任务一：

library(DBI)

library(RMySQL)

killDbConnections()

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

dbSendQuery(con,'SET NAMES utf8') 

rs <- dbSendQuery(con, "SELECT * FROM article WHERE isdone=1")

words = data.frame(word=c(), freq = c())

while (!dbHasCompleted(rs)) {

    chunk <- dbFetch(rs, 10)

    chunk$abstract<-iconv(chunk$abstract,"WINDOWS-1252","UTF-8")

    count=nrow(chunk)

    cnt=1

    while(cnt<=count){

        str = gsub("[[:punct:]]", "", tolower(chunk[cnt,5]))

        temp = as.vector(unlist(strsplit(str, split = " ")))

        temp_len = length(temp)

        cnt2 = 1

        while(cnt2 <= temp_len){

            if(temp[cnt2] %in% words$word){

                words[words$word == temp[cnt2],]$freq = words[words$word == temp[cnt2],]$freq+1

            }

            else{

                words = rbind(words,data.frame(word=c(temp[cnt2]),freq=c(1)))

            }

            cnt2 = cnt2+1

        }

        cnt = cnt +1

    }

}

head(words)

任务二：

library(wordcloud2)

wordcloud2(words[0:1000,])

任务三：

new_words = words[order(words$freq,decreasing=T),]

del_word = c('of','the','and','in','to','a','that','is','for','buy','with','we','are','an','this','these','as','from','which','at','their','have','or','our','its','but','how','be','as','here','on','can','into','data','between','both','also','by')

words2<-words[which(!(words$word %in% del_word)),]

wordcloud2(words2)