Home » Archived » BIRT » Empty flat files
| Empty flat files [message #525651] |
Wed, 07 April 2010 07:32  |
Eclipse User |
|
|
|
Hello,
I have some empty flat files as data sets (sometimes there is no data
there) and I'm getting a "The flatfile data source is not valid"
message. How can I get the empty data set?
By the way, I'm also using a beforeOpen script to parameterize the flat
file name.
Please help. Thanks, Helmut
|
|
| |
| Re: Empty flat files [message #525988 is a reply to message #525804] |
Thu, 08 April 2010 09:13  |
| Eclipse User |
|
|
|
Jason,
thanks for your answer. I think it will help :).
I tried to use headers. Than BIRT doesn't filter this headers. How can I
filter this header using a scripted query? If I don't use scripting
sometimes BIRT displays the header in the preview, sometimes not. I
wasn't able to see the reason, yet. Can you tell me?
Ok, checking the file size is an ugly way, but I will try this. Perhaps
there will be a better way in future.
Helmut
Jason Weathersby schrieb:
> Helmut,
>
> Do you have headers in the flat files that have data? One option would
> be to check the file size in the beforeOpen, then if the size is 0 swap
> in a dummy flat file with data that can be filtered out using a
> dataset/table filter.
>
> importPackage(Packages.java.ip);
> tstfile = new File(yourfullpathfilename);
> if(tstfile.exists()){
> fsize = tstfile.length();
> if( fsize == 0 ){
> //do something;
> }else{
> //do something else;
> }
> }
>
> Jason
>
> Helmut Neubauer wrote:
>> Hello,
>>
>> I have some empty flat files as data sets (sometimes there is no data
>> there) and I'm getting a "The flatfile data source is not valid"
>> message. How can I get the empty data set?
>>
>> By the way, I'm also using a beforeOpen script to parameterize the flat
>> file name.
>>
>> Please help. Thanks, Helmut
|
|
|
| Re: Empty flat files [message #526035 is a reply to message #525988] |
Thu, 08 April 2010 10:50 |
| Eclipse User |
|
|
|
Helmut,
I tried this with BIRT 2.5.2 and only got the error if the headers where
not present. So if you flat files always have a header then you should
not get an error if no data is in the flat file.
Jason
Helmut Neubauer wrote:
> Jason,
>
> thanks for your answer. I think it will help :).
>
> I tried to use headers. Than BIRT doesn't filter this headers. How can I
> filter this header using a scripted query? If I don't use scripting
> sometimes BIRT displays the header in the preview, sometimes not. I
> wasn't able to see the reason, yet. Can you tell me?
>
> Ok, checking the file size is an ugly way, but I will try this. Perhaps
> there will be a better way in future.
>
> Helmut
>
> Jason Weathersby schrieb:
>> Helmut,
>>
>> Do you have headers in the flat files that have data? One option would
>> be to check the file size in the beforeOpen, then if the size is 0 swap
>> in a dummy flat file with data that can be filtered out using a
>> dataset/table filter.
>>
>> importPackage(Packages.java.ip);
>> tstfile = new File(yourfullpathfilename);
>> if(tstfile.exists()){
>> fsize = tstfile.length();
>> if( fsize == 0 ){
>> //do something;
>> }else{
>> //do something else;
>> }
>> }
>>
>> Jason
>>
>> Helmut Neubauer wrote:
>>> Hello,
>>>
>>> I have some empty flat files as data sets (sometimes there is no data
>>> there) and I'm getting a "The flatfile data source is not valid"
>>> message. How can I get the empty data set?
>>>
>>> By the way, I'm also using a beforeOpen script to parameterize the flat
>>> file name.
>>>
>>> Please help. Thanks, Helmut
|
|
|
| Re: Empty flat files (+additional question concerning CSV flat file queries with column names) [message #526561 is a reply to message #526035] |
Mon, 12 April 2010 05:44 |
| Eclipse User |
|
|
|
Jason,
thanks, this works!
First I forgot to check "Use first line as column name indicator" in the
edit data source dialog. So BIRT told me the flat file data source isn't
valid even if I was using the headers. Now all is ok :).
Perhaps one more question, because I use headers the first time in my
CSV files in combination with BIRT: In the past I was scripting my CSV
queries like (found an example in the newsgroups, which I understand
only partially):
queryText= "select \"COLUMN_1\", \"COLUMN_2\" " +
"from " +
params["id"].value + "-1.csv" +
" : {\"COLUMN_1\",\"COLUMN_1\",DATE;\"COLUMN_2\",\"COLUMN_2\ ",INTEGER}";
Now BIRT tells me this isn't ok. I think, because I don't use the
correct names (headers). Where Do I have to use the column names? Can
you explain me how to use the flat file queries?
Thanks,
Helmut
Jason Weathersby schrieb:
> Helmut,
>
> I tried this with BIRT 2.5.2 and only got the error if the headers where
> not present. So if you flat files always have a header then you should
> not get an error if no data is in the flat file.
>
> Jason
>
> Helmut Neubauer wrote:
>> Jason,
>>
>> thanks for your answer. I think it will help :).
>>
>> I tried to use headers. Than BIRT doesn't filter this headers. How can I
>> filter this header using a scripted query? If I don't use scripting
>> sometimes BIRT displays the header in the preview, sometimes not. I
>> wasn't able to see the reason, yet. Can you tell me?
>>
>> Ok, checking the file size is an ugly way, but I will try this. Perhaps
>> there will be a better way in future.
>>
>> Helmut
>>
>> Jason Weathersby schrieb:
>>> Helmut,
>>>
>>> Do you have headers in the flat files that have data? One option would
>>> be to check the file size in the beforeOpen, then if the size is 0 swap
>>> in a dummy flat file with data that can be filtered out using a
>>> dataset/table filter.
>>>
>>> importPackage(Packages.java.ip);
>>> tstfile = new File(yourfullpathfilename);
>>> if(tstfile.exists()){
>>> fsize = tstfile.length();
>>> if( fsize == 0 ){
>>> //do something;
>>> }else{
>>> //do something else;
>>> }
>>> }
>>>
>>> Jason
>>>
>>> Helmut Neubauer wrote:
>>>> Hello,
>>>>
>>>> I have some empty flat files as data sets (sometimes there is no data
>>>> there) and I'm getting a "The flatfile data source is not valid"
>>>> message. How can I get the empty data set?
>>>>
>>>> By the way, I'm also using a beforeOpen script to parameterize the flat
>>>> file name.
>>>>
>>>> Please help. Thanks, Helmut
|
|
|
| Re: Empty flat files (+additional question concerning CSV flat file queries with column names) [message #526696 is a reply to message #526561] |
Mon, 12 April 2010 11:48 |
| Eclipse User |
|
|
|
Helmut
The column names must match what you have in the flat file, unless you
are not using the first line for headers. Ie you should have something
like:
COLUMN_1,COLUMN_2
111,222
..
..
In your flat file.
The query text just has the column names that you are selecting and from
which flatfile. After the : is a list of each column available in the
flat file. These will be in the {} and separated by ;. Each column
will have name alias and type separated by ,.
Jason
Helmut Neubauer wrote:
> Jason,
>
> thanks, this works!
>
> First I forgot to check "Use first line as column name indicator" in the
> edit data source dialog. So BIRT told me the flat file data source isn't
> valid even if I was using the headers. Now all is ok :).
>
> Perhaps one more question, because I use headers the first time in my
> CSV files in combination with BIRT: In the past I was scripting my CSV
> queries like (found an example in the newsgroups, which I understand
> only partially):
>
> queryText= "select \"COLUMN_1\", \"COLUMN_2\" " +
> "from " +
> params["id"].value + "-1.csv" +
> " : {\"COLUMN_1\",\"COLUMN_1\",DATE;\"COLUMN_2\",\"COLUMN_2\ ",INTEGER}";
>
> Now BIRT tells me this isn't ok. I think, because I don't use the
> correct names (headers). Where Do I have to use the column names? Can
> you explain me how to use the flat file queries?
>
> Thanks,
> Helmut
>
> Jason Weathersby schrieb:
>> Helmut,
>>
>> I tried this with BIRT 2.5.2 and only got the error if the headers where
>> not present. So if you flat files always have a header then you should
>> not get an error if no data is in the flat file.
>>
>> Jason
>>
>> Helmut Neubauer wrote:
>>> Jason,
>>>
>>> thanks for your answer. I think it will help :).
>>>
>>> I tried to use headers. Than BIRT doesn't filter this headers. How can I
>>> filter this header using a scripted query? If I don't use scripting
>>> sometimes BIRT displays the header in the preview, sometimes not. I
>>> wasn't able to see the reason, yet. Can you tell me?
>>>
>>> Ok, checking the file size is an ugly way, but I will try this. Perhaps
>>> there will be a better way in future.
>>>
>>> Helmut
>>>
>>> Jason Weathersby schrieb:
>>>> Helmut,
>>>>
>>>> Do you have headers in the flat files that have data? One option would
>>>> be to check the file size in the beforeOpen, then if the size is 0 swap
>>>> in a dummy flat file with data that can be filtered out using a
>>>> dataset/table filter.
>>>>
>>>> importPackage(Packages.java.ip);
>>>> tstfile = new File(yourfullpathfilename);
>>>> if(tstfile.exists()){
>>>> fsize = tstfile.length();
>>>> if( fsize == 0 ){
>>>> //do something;
>>>> }else{
>>>> //do something else;
>>>> }
>>>> }
>>>>
>>>> Jason
>>>>
>>>> Helmut Neubauer wrote:
>>>>> Hello,
>>>>>
>>>>> I have some empty flat files as data sets (sometimes there is no data
>>>>> there) and I'm getting a "The flatfile data source is not valid"
>>>>> message. How can I get the empty data set?
>>>>>
>>>>> By the way, I'm also using a beforeOpen script to parameterize the flat
>>>>> file name.
>>>>>
>>>>> Please help. Thanks, Helmut
|
|
|
Goto Forum:
Current Time: Wed Nov 05 12:00:17 EST 2025
Powered by FUDForum. Page generated in 0.04884 seconds |
Home