I have a portion of code where I pick the value in a button and use it for other purposes. Or, at least, this is what I'd like to do.
The button changes value at every refresh of the page (it's a webpage).
For example: at the first access the button's value (or label) is "Results List (51)" but, if I refresh the page, the value becomes "Results List (11)".
What changes is the number inside the brackets (that identifies the number of results inside the list).
This is the code interested:
ok = Browser("Bwr").Page("Page").Frame("Frame").WebButton("name:=Results List OK").GetToProperty("name")
ko = Browser("Bwr").Page("Page").Frame("Frame").WebButton("name:=Results List KO").GetToProperty("name")
If InStr(ko, "0") > 0 and Instr(ok, "0")=0 Then
reporter.ReportEvent 0, "Riabbinamento effettuato", "Operazione effettuata con esito positivo: tutte le misure sono state riabbinate"
else reporter.ReportEvent 1, "Riabbinamento fallito", "Operazione effettuata con esito negativo: ci sono misure su cui l'operazione è fallita"
End If
Don't pay attention to the reporter (I'm Italian, it's written in my language).
If I execute the above code QTP puts in ok the string "Results List OK" but I want to put in ok the string "Results List OK (n)" (with n being the number that changes at every refresh of the page).
Basically I only need the number inside the brackets in order to make the IF truly works...
Any idea?!
You want to use a regular expression to map the property.
Result List (\d+)
or just Result List.*
Ok problem solved.
I've used GetRoProperty instead of GetToProperty and modified the value in the brackets after WebElement from "name:=Results List OK" to "name:=Results List OK.*"
Thanks to gigatropolis for the useful tips (I upvoted your answer) but it was only half the solution :)
Related
Anybody have any tips for looping, and continue? For example, I placed about 2500 pnrs on a queue, and I need to add a remark to each of them. Is it possible for a script to add the remark then move to the next pnr?
For example, I placed about 2500 pnrs on a queue, and I need to add a remark to each of them. Is it possible for a script to add the remark then move to the next pnr?
Loops are supported in Scribe but have to be built manually by creating your own iteration variables and breaking the loop manually when you know the work is complete.
What you are describing is definitely possible, but working in queues can be difficult as there are many possible responses when you try to end the PNRs. You'll have to capture the response to detect whether you need to do something else to get out of the error condition (e.g. if a PNR warning indicates you have to double-end the record).
If possible, its likely simpler to work off the queue by collecting all PNR locators and then looping through that list, adding your remarks, and then ending the PNRs. You'll still have to capture the response to determine if the PNR is actually ended properly, but you won't have to deal with the buggy queue behavior. A basic Scribe loop sample is below. Note that I haven't been a Scribe developer for a while and I did this in Notepad so there might be some errors in here, but hopefully it's a good starting point.
DEFINE [ROW=N:8] ;iteration variable/counter
DEFINE [LOCATOR_FILE=*:60] ;File Path
DEFINE [TEMP_LOCATOR=*:6] ;pnr locator variable, read from the temp file
DEFINE [BREAK=*:1] ;loop breaking variable
OPEN F=[TEMP_LOCATOR] L=0 ;open the file of locators
[BREAK] = ""
[ROW] = 0
REPEAT
[ROW] = [ROW] + 1
[TEMP_LOCATOR] = "" ;Reset temp locator variable, this will break our loop
READ F=[LOCATOR_FILE] R=[ROW] C=1 [TEMP_LOCATOR]
IF $[TEMP_LOCATOR] = 6 THEN ;test length of locator, if this is 6 chars, you have a good one, pull it up and add your remark
»"5YOUR REMARK HERE"{ENTER}«
»ER{ENTER}«
;trap errors
READ F="EMUFIND:" R=0 C=0 [TEMP_LOCATOR] ;read for the locator being present on this screen, which should indicate that the ER was successful - you'll have to trap other errors here though
IF [#SYSTEM_ERROR] = 0 THEN ;this locator was found, ER appears successful
»I{ENTER}« ;Ignore this PNR and move to the next one
ELSE
[BREAK] = "Y" ;error found afeter ER, break loop. Maybe show a popup box or something, up to you
ENDIF
ELSE ;No locator found in file, break the loop
[BREAK] = "Y"
ENDIF
UNTIL [BREAK] = "Y"
CLOSE [LOCATOR_FILE]
I have a webpage with a combobox to sort a list by Price, Name,... The list is in many pages, so I need to get all elements first and after, I will sort by and check if the elements are correct. Right?
I am trying to do it and navigate for all pages and get all elements. But it is only taking the elements in first page. I am totally new in Robot framework.
Does anyone have a suggestion how I can do it?
${name_list_actual} Create List
${cnt_pages}= Get Element Count //div[#class='container index-new-p']/div/nav[#class='text-center']/ul/li/*
Log To Console ${cnt_pages}
:FOR ${n} IN RANGE 1 ${cnt_pages}
\ Click link //div[#class='container index-new-p']/div/nav/ul/li[${n}]/a
\ ${cnt}= Get Element Count //*[#class="title"]/*
:FOR ${i} IN RANGE 1 ${cnt}
\ ${get_names} Get Element Attribute //table[#class='result-table']/tbody/tr[${i}]/td/div/div[2]/div/a text
\ Append To List ${name_list_actual} ${get_names}
${get_names}= Select All From List //table[#class='result-table']/tbody/tr[${i}]/td/div/div[2]/div/a
\ Log To Console ${name_list_actual}
\ Continue For Loop
Thanks so much
First and foremost, by your code sample's indent it looks like you're doing nested for loops - an outer one for the changing the pages, and an inner that'll get the text in each row. The thing is - Robot Framework doesn't support nested for loops - see the documentation.
What happened on execution is the first loop (the pagination) ran with just two statements and finished:
\ Click link //div[#class='container index-new-p']/div/nav/ul/li[${n}]/a
\ ${cnt}= Get Element Count //*[#class="title"]/*
Then all the rest were executed as part of the second.
A solution would be to have a keyword "Get All Values In The Table" that'll hold the inner loop, and call it in the current outer, which will paginate.
Another issue - you're using IN RANGE, passing 1 as start and ${cnt} as end; if it had value of 4, you probably expect the tracking variable to get all values from 1 to 4.
Yet range works with the first argument inclusive, and up to - but not getting to - the second argument; thus it covers the range 1 to 3. To salvage that, you'd need to set the upper bound to cnt+1, e.g. ${cnt + 1}.
Minor stuff:
You have a statement that's within the loop block, but not prefixed with \:
${get_names}= Select All From List //table[#class='result-table']/tbody/tr[${i}]/td/div/div[2]/div/a
I'm surprised it didn't give you a syntax error - it effectively breaks the loop, and the framework should complain the next lines are prefixed as if in a loop, but one is not started.
-
When you are changing the pages by the click, you are not making sure the UI has loaded the data for the new page. If this happens through ajax calls, you may very well be working with the previous page's data, thinking it is the new one - the selenium click returns control very fast, and the operands for getting the text are running - while the UI still waits for the new data and is displaying the previous one.
-
In RF version 3.1 the loop syntax is different - the block members are not prefixed with \, and it's closed with an END statement.
I'd suggest to migrate to it - a) the current one is going to be eventually deprecated, and b) in the new one the nesting would have been marked as syntax error (you're starting a new loop without END-ing the running one), instead of silently passing with unexpected behavior.
I am following my question from yesterday - harvesting data via drop down list in R 1
first, I need to obtain all 50k strings of details of all doctors from this page: http://www.lkcr.cz/seznam-lekaru-426.html#seznam
I know, how to obtain them from a single page:
oborID<-"48"
okresID<-"3702"
web<- "http://www.lkcr.cz/seznam-lekaru-426.html"
extractHTML<-function(oborID,okresID){
query<-list('filterObor'="107",'filterOkresId'="3201",'do[findLekar]'=1)
query$filterObor<-oborID
query$filterOkresId<-okresID
html<- POST(url=web,body=query)
html<- content(html, "text")
html
}
IDfromHTML<-function(html){
starting<- unlist(gregexpr("filterId", html))
ending<- unlist(gregexpr("DETAIL", html))
starting<- starting[seq(2,length(starting),2)]
if (starting != -1 && ending != -1){
strings<-c()
for (i in 1:length(starting)) {
strings[i]<-substr(html,starting[i]+9,ending[i]-18)
}
strings<-list(strings)
strings
}
}
still, I am aware that downloading whole page for only few lines of text is quite uneffective(but works!:) Could you give me a tip how to make this process more effective?
I have also encountered some pages with more than 20 doctors listed (i.e. combination of "Brno-město" and "chirurgie". Such data are listed and accessed via hyperlink list at the end of the form. I need to access each of these pages and use there the code I presented here. But I guess I have to pass some cookies there.
Other than that, combination of "Praha" and "chirurgie" is problematic as well, because there is more than 200 records, therefore page applies some script and then I need to click the button "další" and use the same method as in the previous paragraph.
Can you help me please?
I am searching for a way to get user input inside a loop while executing in batch mode.
readLines() and scan() work well for me in interactive mode only, in batch mode they start to read in lines of code as user input, unless all the code is surrounded by {}, which is inconvenient. I need a simple solution to get just 1 integer value in a way that I can just type in value and press ENTER, so
the input field (if the solution involves GUI) must automatically get focus and
ENTER must trigger end of input/submission.
I can't find a way to do it that will satisfy both conditions, e.g. ginput() from gWidgets activates the input field, but ENTER doesn't trigger form submission.
Here is how I solved my own problem:
require(gWidgets)
options(guiToolkit="RGtk2")
INPUT <- function(message) {
CHOICE <- NA
w <- gbasicdialog(title=message, handler = function(h,...) CHOICE <<- svalue(input))
input <- gedit("", initial.msg="", cont=w, width=10)
addHandlerChanged(input, handler=function (h, ...) {
CHOICE <<- svalue(input)
dispose(w)
})
visible(w, set=TRUE)
return(CHOICE)
}
repeat{
x=INPUT("Input an integer")
if(!is.na(as.integer(x))) break
}
print(x)
Update:
I can't test this right now, but take a look at ?menu and have it pop up a gui window.
I'm not certain if that will work, but it is different in that it takes a mouse-click response.
original answer:
As per the documentation to ?readline:
This can only be used in an interactive session.
..
In non-interactive use the result is as if the response was RETURN and the value is "".
If you are simply waiting for one piece of information, and you do not know this piece of information before beginning the execution of the script (presumably, there is a decision to be made which is dependent on the results earlier in the script), then one alternative is to simply break your script up into three parts:
everything before the decision point.
an interactive script which prompts for input
everything after the decision point.
And simply chain the three together by having the first end by calling the second in an interactive session. Then have the second end by calling the third.
(first question here, sorry if I am breaking a piece of etiquette)
My site is running on an eCommerce back end provider that I subscribe to. They have everything in classic ASP. They have a black box function called import_products that I use to import a given text file into my site's database.
The problem is that if I call the function more than once, something breaks. Here is my example code:
for blah = 1 to 20
thisfilename = "fullcatalog_" & blah & ".csv"
Response.Write thisfilename & "<br>"
Response.Flush
Call Import_Products(3,thisfilename,1)
Next
Response.End
The first execution of the Import_Products function works fine. The second time I get:
Microsoft VBScript runtime error '800a0009'
Subscript out of range: 'i'
The filenames all exist. That part is fine. There are no bugs in my calling code. I have tried checking the value of "i" before each execution. The first time the value is blank, and before the second execution the value is "2". So I tried setting it to null during each loop iteration, but that didn't change the results at all.
I assume that the function is setting a variable or opening a connection during its execution, but not cleaning it up, and then not expecting it to already be set the second time. Is there any way to find out what this would be? Or somehow reset the condition back to nothing so that the function will be 'fresh'?
The function is in an unreadable include file so I can't see the code. Obviously a better solution would be to go with the company support, and I have a ticket it in with them, but it is like pulling teeth to get them to even acknowledge that there is a problem. Let alone solve it.
Thanks!
EDIT: Here is a further simplified example of calling the function. The first call works. The second call fails with the same error as above.
thisfilename = "fullcatalog_testfile.csv"
Call Import_Products(3,thisfilename,1)
Call Import_Products(3,thisfilename,1)
Response.End
The likely cause of the error are the two numeric parameters for the Import_Products subroutine.
Import_Products(???, FileName, ???)
The values are 3 and 1 in your example but you never explain what they do or what they are documented to do.
EDIT Since correcting the vender subroutine is impossible, but it always works for the first time it's called lets use an HTTP REDIRECT instead of a FOR LOOP so that it technically only gets called once per page execution.
www.mysite.tld/import.asp?current=1&end=20
curr = CInt(Request.QueryString("current"))
end = CInt(Request.QueryString("end"))
If curr <= end Then
thisfilename = "fullcatalog_" & curr & ".csv"
Call Import_Products(3,thisfilename,1)
Response.Redirect("www.mysite.tld/import.asp?current=" & (curr + 1) & "&end=" & end)
End If
note the above was written inside my browser and is untested so syntax errors may exist.