Cross Table with Embedded Labels AND explicit

[shane_spencer]

Is it possible to use a combination of Embedded Labels AND explicit for use on a cross table?

For example I've got 3 types of data file that I want to load in to Qlikview.

Column 1 and Column 2 are Qualifier Fields.

Column 1 header is always [00:00:01] and will be loaded as "Time",

Column 2 header changes from file to file and will be loaded as "Main_metric"

...so ideally I want to use Explicit values.

Column 3 and beyond are Attribute Fields and Data Fields.

The header in the raw data is needed for the Attribute Field so Imbedded labels are needed.

Ideally I want to replace 3 load scripts with 1, is that possible?

i.e. replace:

CrossTable(sub_Metric, value, 2)

LOAD [00:00:01] as Time,

     IFACE as Main_metric,

     [rxerr/s],

     [txerr/s],

     [coll/s],

     [rxdrop/s],

     [txdrop/s],

     [txcarr/s],

     [rxfram/s],

     [rxfifo/s],

     [txfifo/s]

CrossTable(sub_Metric, value, 2)

LOAD [00:00:01] as Time,

     CPU as Main_metric,

     %usr,

     %nice,

     %sys,

     %iowait,

     %steal,

     %irq,

     %soft,

     %guest,

     %idle

CrossTable(sub_Metric, value, 2)

LOAD [00:00:01] as Time,

     DEV as Main_metric,

     tps,

     [rd_sec/s],

     [wr_sec/s],

     [avgrq-sz],

     [avgqu-sz],

     await,

     svctm,

     %util

...with a single load statement something like :

CrossTable(sub_Metric, value, 2)

LOAD [00:00:01] as Time,

     @2 as Main_metric, 

     *

[marcus_sommer · Accepted solution]

A mix of embedded label and no label isn't possible with txt-files. But you could easily load this few fields as column-no - see attachment.

- Marcus

[shane_spencer]

Hi Marcus, Thanx, but I have 5 files (with 3 different Headers in Column 2, and all 5 have different headers in the subsequent columns) so I would be more efficient to use embedded labels and have 3 load scripts rather than have 5 load scripts without embedded labels.

I suspect what I want to do is not possible, but I wondered if perhaps it could be done by maybe creating an empty table then loading data in to it.

[marcus_sommer]

I wouldn't say that is impossible but definitely not useful - different is different and every difference needed to be handled with some condition (perhaps in loops with variables and/or the rename-statement). But in your case with only a few tables/fields it would be faster and better to keep the script simple.

- Marcus

[shane_spencer]

Cheers Marcus - I was hoping because the files were so similar that there'd be an easy / graceful work-around. At present I can just do it with 3 load scripts but I've no idea if in the future there'll be more similar files to process that will require more scripts.