html parsing of cricinfo scorecards - html

Question

Aim
I am looking to scrape 20/20 cricket scorecard data from the Cricinfo website, ideally into CSV form for data analysis in Excel
As an example the current Australian Big Bash 2011/12 scorecards are available from
Game 1: http://www.espncricinfo.com/big-bash-league-2011/engine/match/524915.html
Last Game: http://www.espncricinfo.com/big-bash-league-2011/engine/match/524935.html
Background
I am proficient in using VBA (either automating IE or using XMLHTTP and then using regular expressions) to scrape data from websites, ie
Extract values from HTML TD and Tr
In that same question a comment was posted suggesting html parsing - which I hadn't come accross before - so I have taken a look at questions such as RegEx match open tags except XHTML self-contained tags
Query
While I could write a regex to parse the cricket data below I would like advice as to how I could efficiently retrieve these results with html parsing.
Please bear in mind that my preference is a repeatable CSV format containing:
the date/name of the match
Team 1 name
the output should dump up to 11 records for Team 1 (blank records where players haven't batted, ie "Did Not Bat")
Team 2 name
the output should dump up to 11 records for Team 2 (blank records where players haven't batted)
Nirvana for me would be a solution that I could deploy using VBA or VBscript so I could fully automate my analysis, but I presume I will have to use a separate tool for the html parse.
Sample Site links and Data to be Extracted

Answer 1

There are 2 techniques that I use for "VBA". I will describe them 1 by one.
1) Using FireFox / Firebug Addon / Fiddler
2) Using Excel's inbuilt facility to get data from the web
Since this post will be read by many so I will even cover the obvious. Please feel free to skip whatever part you know
1) Using FireFox / Firebug Addon / Fiddler
FireFox : http://en.wikipedia.org/wiki/Firefox
Free download (http://www.mozilla.org/en-US/firefox/new/)
Firebug Addon: http://en.wikipedia.org/wiki/Firebug_%28software%29
Free download (https://addons.mozilla.org/en-US/firefox/addon/firebug/)
Fiddler : http://en.wikipedia.org/wiki/Fiddler_%28software%29
Free download (http://www.fiddler2.com/fiddler2/)
Once you have installed Firefox, install the Firebug Addon. The Firebug Addon lets you inspect the different elements in a webpage. For example if you want to know the name of a button, simply right click on it and click on "Inspect Element with Firebug" and it will give you all the details that you will need for that button.
Another example would be finding the name of a table on a website which has the data that you need scrapped.
I use Fiddler only when I am using XMLHTTP. It helps me to see the exact info being passed when you click on a button. Because of the increase in the number of BOTS which scrape the sites, most sites now, to prevent automatic scrapping, capture your mouse coordinates and pass that information and fiddler actually helps you in debugging that info that is being passed. I will not get into much details here about it as this info can be used maliciously.
Now let's take a simple example on how to scrape the URL posted in your question
http://www.espncricinfo.com/big-bash-league-2011/engine/match/524915.html
First let's find the name of the table which has that info. Simply right click on the table and click on "Inspect Element with Firebug" and it will give you the below snapshot.
So now we know that our data is stored in a table called "inningsBat1" If we can extract the contents of that table to an Excel file then we can definitely work with the data to do our analysis. Here is sample code which will dump that table in Sheet1
Before we proceed, I would recommend, closing all Excel and starting a fresh instance.
Launch VBA and insert a Userform. Place a command button and a webcrowser control. Your Userform might look like this
Paste this code in the Userform code area
Option Explicit
'~~> Set Reference to Microsoft HTML Object Library
Private Declare Sub Sleep Lib "kernel32" (ByVal dwMilliseconds As Long)
Private Sub CommandButton1_Click()
Dim URL As String
Dim oSheet As Worksheet
Set oSheet = Sheets("Sheet1")
URL = "http://www.espncricinfo.com/big-bash-league-2011/engine/match/524915.html"
PopulateDataSheets oSheet, URL
MsgBox "Data Scrapped. Please check " & oSheet.Name
End Sub
Public Sub PopulateDataSheets(wsk As Worksheet, URL As String)
Dim tbl As HTMLTable
Dim tr As HTMLTableRow
Dim insertRow As Long, Row As Long, col As Long
On Error GoTo whoa
WebBrowser1.navigate URL
WaitForWBReady
Set tbl = WebBrowser1.Document.getElementById("inningsBat1")
With wsk
.Cells.Clear
insertRow = 0
For Row = 0 To tbl.Rows.Length - 1
Set tr = tbl.Rows(Row)
If Trim(tr.innerText) <> "" Then
If tr.Cells.Length > 2 Then
If tr.Cells(1).innerText <> "Total" Then
insertRow = insertRow + 1
For col = 0 To tr.Cells.Length - 1
.Cells(insertRow, col + 1) = tr.Cells(col).innerText
Next
End If
End If
End If
Next
End With
whoa:
Unload Me
End Sub
Private Sub Wait(ByVal nSec As Long)
nSec = nSec + Timer
While Timer < nSec
DoEvents
Sleep 100
Wend
End Sub
Private Sub WaitForWBReady()
Wait 1
While WebBrowser1.ReadyState <> 4
Wait 3
Wend
End Sub
Now run your Userform and click on the Command button. You will notice that the data is dumped in Sheet1. See snapshot
Similarly you can scrape other info as well.
2) Using Excel's inbuilt facility to get data from the web
I believe you are using Excel 2007 so I will take that as an example to scrape the above mentioned link.
Navigate to Sheet2. Now navigate to Data Tab and click on the button "From Web" on the extreme right. See snapshot.
Enter the url in the "New Web Query Window" and click on "Go"
Once the page is uploaded, select the relevant table that you want to import by clicking on the small arrow as shown in the snapshot. Once done, click on "Import"
Excel will then ask you where you want the data to be imported. Select the relevant cell and click on OK. And you are done! The data will be imported to the cell which you specified.
If you wish you can record a macro and automate this as well :)
Here is the macro that I recorded.
Sub Macro1()
With ActiveSheet.QueryTables.Add(Connection:= _
"URL;http://www.espncricinfo.com/big-bash-league-2011/engine/match/524915.html" _
, Destination:=Range("$A$1"))
.Name = "524915"
.FieldNames = True
.RowNumbers = False
.FillAdjacentFormulas = False
.PreserveFormatting = True
.RefreshOnFileOpen = False
.BackgroundQuery = True
.RefreshStyle = xlInsertDeleteCells
.SavePassword = False
.SaveData = True
.AdjustColumnWidth = True
.RefreshPeriod = 0
.WebSelectionType = xlSpecifiedTables
.WebFormatting = xlWebFormattingNone
.WebTables = """inningsBat1"""
.WebPreFormattedTextToColumns = True
.WebConsecutiveDelimitersAsOne = True
.WebSingleBlockTextImport = False
.WebDisableDateRecognition = False
.WebDisableRedirections = False
.Refresh BackgroundQuery:=False
End With
End Sub
Hope this helps. Let me know if you still have some queries.
Sid

Answer 2

For anyone else interested in this I ended up using the code below based on Siddhart Rout's earlier answer
XMLHttp was significantly quicker than automating IE
the code generates a CSV file for each series to be dowloaded (held in the X variable)
the code dumps each match to a regular 29 row range (regardless of how many players batted) to facillitate easier analysis later on
Public Sub PopulateDataSheets_XML()
Dim URL As String
Dim ws As Worksheet
Dim lngRow As Long
Dim lngRecords As Long
Dim lngWrite As Long
Dim lngSpare As Long
Dim lngInnings As Long
Dim lngRow1 As Long
Dim X(1 To 15, 1 To 4) As String
Dim objFSO As Object
Dim objTF As Object
Dim xmlHttp As Object
Dim htmldoc As HTMLDocument
Dim htmlbody As htmlbody
Dim tbl As HTMLTable
Dim tr As HTMLTableRow
Dim strInnings As String
s = Timer()
Set xmlHttp = CreateObject("MSXML2.ServerXMLHTTP")
Set objFSO = CreateObject("scripting.filesystemobject")
X(1, 1) = "http://www.espncricinfo.com/indian-premier-league-2011/engine/match/"
X(1, 2) = 501198
X(1, 3) = 501271
X(1, 4) = "indian-premier-league-2011"
X(2, 1) = "http://www.espncricinfo.com/big-bash-league-2011/engine/match/"
X(2, 2) = 524915
X(2, 3) = 524945
X(2, 4) = "big-bash-league-2011"
X(3, 1) = "http://www.espncricinfo.com/ausdomestic-2010/engine/match/"
X(3, 2) = 461028
X(3, 3) = 461047
X(3, 4) = "big-bash-league-2010"
Set htmldoc = New HTMLDocument
Set htmlbody = htmldoc.body
For lngRow = 1 To UBound(X, 1)
If Len(X(lngRow, 1)) = 0 Then Exit For
Set objTF = objFSO.createtextfile("c:\temp\" & X(lngRow, 4) & ".csv")
For lngRecords = X(lngRow, 2) To X(lngRow, 3)
URL = X(lngRow, 1) & lngRecords & ".html"
xmlHttp.Open "GET", URL
xmlHttp.send
Do While xmlHttp.Status <> 200
DoEvents
Loop
htmlbody.innerHTML = xmlHttp.responseText
objTF.writeline X(lngRow, 1) & lngRecords & ".html"
For lngInnings = 1 To 2
strInnings = "Innings " & lngInnings
objTF.writeline strInnings
Set tbl = Nothing
On Error Resume Next
Set tbl = htmlbody.Document.getElementById("inningsBat" & lngInnings)
On Error GoTo 0
If Not tbl Is Nothing Then
lngWrite = 0
For lngRow1 = 0 To tbl.Rows.Length - 1
Set tr = tbl.Rows(lngRow1)
If Trim(tr.innerText) <> vbNewLine Then
If tr.Cells.Length > 2 Then
If tr.Cells(1).innerText <> "Extras" Then
If Len(tr.Cells(1).innerText) > 0 Then
objTF.writeline strInnings & "-" & lngWrite & "," & Trim(tr.Cells(1).innerText) & "," & Trim(tr.Cells(3).innerText)
lngWrite = lngWrite + 1
End If
Else
objTF.writeline strInnings & "-" & lngWrite & "," & Trim(tr.Cells(1).innerText) & "," & Trim(tr.Cells(3).innerText)
lngWrite = lngWrite + 1
Exit For
End If
End If
End If
Next
For lngSpare = 12 To lngWrite Step -1
objTF.writeline strInnings & "-" & lngWrite + (12 - lngSpare)
Next
Else
For lngSpare = 1 To 13
objTF.writeline strInnings & "-" & lngWrite + (12 - lngSpare)
Next
End If
Next
Next
Next
'Call ConsolidateSheets
End Sub

Answer 3

RegEx is not a complete solution for parsing HTML because it is not guaranteed to be regular.
You should use the HtmlAgilityPack to query the HTML. This will allow you to use the CSS selectors to query the HTML similar to how you do it with jQuery.

Answer 4

As quite a few people may see this I thought I would use it as a chance to demonstrate a few features I rarely see people using in VBA web-scraping: deleteRow, querySelector and use of clipboard to write out a table (complete with formatting and hyperlinks) to a sheet based on the table.outerHTML.
deleteRow is used to remove the unwanted rows. querySelector is used to apply faster css selectors to match on nodes. Modern browsers/html parsers are optimized for css and class selectors (which I use) are the second fastest selector type (after id).
Use of css selectors and understanding htmlTable methods/properties will allow for much greater flexibility in your web-scraping endeavours. Understanding the use of the clipboard means a simple copy paste method for transferring a table to Excel.
Execution could easily be tied to a button push and the url read in from a cell.
VBA:
Option Explicit
Public Sub test()
WriteOutTable "https://www.espncricinfo.com/series/8044/scorecard/524935/hobart-hurricanes-vs-melbourne-stars-big-bash-league-2011-12"
End Sub
Public Sub WriteOutTable(ByVal url As String)
'required VBE (Alt+F11) > Tools > References > Microsoft HTML Object Library ; Microsoft XML, v6 (your version may vary)
Dim hTable As MSHTML.HTMLTable, clipboard As Object
Dim xhr As MSXML2.xmlhttp60, html As MSHTML.htmlDocument
Set xhr = New MSXML2.xmlhttp60
Set html = New MSHTML.htmlDocument
With xhr
.Open "GET", url, False
.Send
html.body.innerHTML = .responseText
End With
Set hTable = html.querySelector(".batsman")
rowCount = hTable.Rows.Length - 1
For i = rowCount To 0 Step -1
Select Case True
Case i = rowCount Or i = rowCount - 1 Or InStr(hTable.Rows(i).outerHTML, "wicket-details") > 0
hTable.deleteRow i
End Select
Next
Set clipboard = GetObject("New:{1C3B4210-F441-11CE-B9EA-00AA006B1A69}")
clipboard.SetText hTable.outerHTML
clipboard.PutInClipboard
ActiveSheet.Cells(1, 1).PasteSpecial
End Sub