HTTP_Resource_Services

BarryStevens · July 2016

I am testing the return of only master dict columns, but I am getting all the columns.
Here is the code extract from the last copy I have.
this is and extract from the GetColumnNames paragraph.

If you look at the line:
If Len(ColumnNames) else ColumnNames = Fields

Fields is all the column names.
Shouldn't it be using an array from where the delete(...... is done?

I know I am probably being stupid as usual, but I cant see it working.

Loop
                       Master      = Fields<FIELDS_MASTER_FLAG$, FieldPos>
                       Type        = Fields<FIELDS_TYPE$, FieldPos>
                       Conv        = Fields<FIELDS_CONV$, FieldPos>
                       MV          = Fields<FIELDS_MVFLAG$, FieldPos>
                       Pos         = Fields<FIELDS_FIELD_NO$, FieldPos>
                       KeyPart     = Fields<FIELDS_PART$, FieldPos>
                       GroupName   = Fields<FIELDS_MV_GROUPNAME$, FieldPos>
 
                       If (Master AND Not(Pos EQ 0 AND KeyPart EQ 0)) OR (Type EQ 'S') then
                           // The column must be a Master column (but not a full Key ID) or a Calculated column.
                           If Fields<FIELDS_NAME$, FieldPos> EQ '' then
                               // If the column does not have a name in the FIELDS_NAME$ value (for whatever reason) then
                               // remove it from the array so it will be be included in the final results.
                               Fields      = Delete(Fields, FIELDS_FIELD_NO$, FieldPos, 0)
                               Fields      = Delete(Fields, FIELDS_PART$, FieldPos, 0)
                               FieldsList  = Delete(FieldsList, FieldPos, 0, 0)
                           end
                       end else
                           Fields      = Delete(Fields, FIELDS_FIELD_NO$, FieldPos, 0)
                           Fields      = Delete(Fields, FIELDS_PART$, FieldPos, 0)
                           FieldsList  = Delete(FieldsList, FieldPos, 0, 0)
                           Locate Fields<FIELDS_NAME$, FieldPos> in ColumnNames using @FM setting fPos then ColumnNames = Delete(ColumnNames, fPos, 0, 0)
                       end
                   Until Fields<FIELDS_FIELD_NO$, FieldPos> GT 0 OR (FieldPos EQ 0)
                       FieldPos -= 1
                   Repeat
                   Fields = SRP_Rotate_Array(FieldsList)
                   If Len(ColumnNames) else ColumnNames = Fields<FIELDS_NAME$>

BarryStevens · July 2016

Ok, Whoops, sorry missed this:
Fields = SRP_Rotate_Array(FieldsList)

Ok MY problem must be elsewhere

BarryStevens · July 2016

I am really sorry if I am still missing something, but the only way I could get the correct result is to change this:

*Until Fields GT 0 OR (FieldPos EQ 0)
Until (FieldPos EQ 0)

Without the change, would not it be exiting on the first field position found?

BarryStevens · July 2016

Sorry, the change should be this:

FieldPos -= 1
Until (FieldPos EQ 0)
repeat

DonBakke · July 2016

Barry,

I really am not at all clear on what you are asking, assuming you still have a question any more. If you still have an unanswered question please post again as simply as possible.

BarryStevens · July 2016

The following code extract is supposed to loop through all the dict fields and create a list of Master fields.
This was not happening for me.

I see the problem in here:

Until Fields GT 0 OR (FieldPos EQ 0)
FieldPos -= 1
Repeat

When the field No is >1 the loop is exited.

If I change to:

FieldPos -= 1
Until (FieldPos EQ 0)
repeat

It works as I supposed it should.

All I am doing is letting you know in case there is a problem with your original code.

GetColumnNames paragraph

Loop
                       Master      = Fields<FIELDS_MASTER_FLAG$, FieldPos>
                       Type        = Fields<FIELDS_TYPE$, FieldPos>
                       Conv        = Fields<FIELDS_CONV$, FieldPos>
                       MV          = Fields<FIELDS_MVFLAG$, FieldPos>
                       Pos         = Fields<FIELDS_FIELD_NO$, FieldPos>
                       KeyPart     = Fields<FIELDS_PART$, FieldPos>
                       GroupName   = Fields<FIELDS_MV_GROUPNAME$, FieldPos>
  
                       If (Master AND Not(Pos EQ 0 AND KeyPart EQ 0)) OR (Type EQ 'S') then
                           // The column must be a Master column (but not a full Key ID) or a Calculated column.
                           If Fields<FIELDS_NAME$, FieldPos> EQ '' then
                               // If the column does not have a name in the FIELDS_NAME$ value (for whatever reason) then
                               // remove it from the array so it will be be included in the final results.
                               Fields      = Delete(Fields, FIELDS_FIELD_NO$, FieldPos, 0)
                               Fields      = Delete(Fields, FIELDS_PART$, FieldPos, 0)
                               FieldsList  = Delete(FieldsList, FieldPos, 0, 0)
                           end
                       end else
                           Fields      = Delete(Fields, FIELDS_FIELD_NO$, FieldPos, 0)
                           Fields      = Delete(Fields, FIELDS_PART$, FieldPos, 0)
                           FieldsList  = Delete(FieldsList, FieldPos, 0, 0)
                           Locate Fields<FIELDS_NAME$, FieldPos> in ColumnNames using @FM setting fPos then ColumnNames = Delete(ColumnNames, fPos, 0, 0)
                       end
                   Until Fields<FIELDS_FIELD_NO$, FieldPos> GT 0 OR (FieldPos EQ 0)
                       FieldPos -= 1
                   Repeat
                   Fields = SRP_Rotate_Array(FieldsList)
                   If Len(ColumnNames) else ColumnNames = Fields<FIELDS_NAME$></div>

DonBakke · July 2016

Barry,

I think some of my confusion lies in the smaller snippets of code you are putting in your post. Are those exact copies or did you simply type it out? I simply don't see any code that looks exactly like that.

The larger snippet of code, of course, comes right from the service. With regard to the placement of the FieldPos -= 1 line, it seems like it would be more efficient if it was placed above the Until statement. However, I don't see where the code would fail to provide you with all of your column names. If anything, the loop goes through one more time than it should.

DonBakke · July 2016

Barry,

Upon reviewing the code again, the original placement of FieldPos -= 1 is by design to properly satisfy the Fields< FIELDS_FIELD_NO$, FieldPos > GT 0 condition. So, perhaps your dictionary poses a unique challenge to the design. If you are willing to ship me a copy of your %FIELDS%, I would be happy to test it.

BarryStevens · July 2016

This is a copy/paste from what I thought was the latest code from you.
Can you send me the latest you have please and I will take a relook.

BarryStevens · July 2016

copy of %FIELDS% sent via email

BarryStevens · July 2016

this is showing as my http frameworks version:
2.1.0
04/11/2016 10:15PM

DonBakke · July 2016

For those monitoring this thread, thanks to Barry's discovery and his %FIELDS% record (which was rather full of all kinds of column goodness), I was able to confirm the issue he brought up. I have updated the GetColumnNames service and it will be released in a new HTTP_Resource_Services module in an upcoming upgrade to the HTTP Framework. There is no scheduled date for this so if anybody just wants a fix for this service, send an email to support@srpcs.com and I'll ship you the source.

HTTP_Resource_Services

Comments