Error during export objectlogs

[bmodeco]

I am logging several values in the objectlogs tab. When I try to export them I get the following error in a new safari-tab:
Error: /lib/genohm-scada/web/objectlogs/csv.lua:10: unexpected symbol near '='
Any ideas how to resolve this?
Thanks, Bart

[Erwin van der Zwart]

Hi Bart,

Not sure, but it looks like you are using special sign in you object names and that might cause a error in the CSV notation.. are you using ' or '' signs or , in your object names?

BR,

Erwin

[admin]

It's a bug in the FW, we will fix it in the next release.

[bmodeco]

Hi Bart,

Not sure, but it looks like you are using special sign in you object names and that might cause a error in the CSV notation.. are you using ' or '' signs or , in your object names?

BR,

Erwin

Hi Erwin,
Only minus-sign nothing else.
I have attached a screendump.
Mvg, Bart

[bmodeco]

It's a bug in the FW, we will fix it in the next release.

Hi Admin,
Any ideas how to get a logfile (csv) for a set of grpadresses?
I would need a 2 day logging of the weatherstation since it is giving wrong values (it is currently more than 50 degrees C in Belgium according to the station ) - need to send it to the supplier.
Best regards,
Bart

[Erwin van der Zwart]

Hi Bart,

This script creates a csv of the object logs on your ftp folder:

Code:

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657

-- get past year data logtime = os.time() - 60 * 60 * 24 * 365 -- list of objects by id objects = {} -- objects with logging enabled query = 'SELECT address, datatype, name FROM objects WHERE disablelog=0' for _, object in ipairs(db:getall(query)) do  objects[ tonumber(object.address) ] = {    datatype = tonumber(object.datatype),    name = tostring(object.name or ''),  } end -- csv buffer buffer = { '"date","address","name","value"' } -- get object logs query = 'SELECT src, address, datahex, logtime, eventtype FROM objectlog WHERE logtime >= ? ORDER BY id DESC' for _, row in ipairs(db:getall(query, logtime)) do  object = objects[ tonumber(row.address) ]  -- found matching object and event type is group write  if object and row.eventtype == 'write' then    datatype = object.datatype    -- check that object datatype is set    if datatype then      -- decode data      data = knxdatatype.decode(row.datahex, datatype)      -- remove null chars from char/string datatype      if datatype == dt.char or datatype == dt.string then        data = data:gsub('%z+', '')      -- date to DD.MM.YYYY      elseif datatype == dt.date then        data = string.format('%.2d.%.2d.%.2d', data.day, data.month, data.year)      -- time to HH:MM:SS      elseif datatype == dt.time then        data = string.format('%.2d:%.2d:%.2d', data.hour, data.minute, data.second)      end    else      data = ''    end    -- format csv row    logdate = os.date('%Y.%m.%d %H:%M:%S', row.logtime)    csv = string.format('%q,%q,%q,%q', logdate, knxlib.decodega(row.address), object.name, tostring(data))    -- add to buffer    table.insert(buffer, csv)  end end dst = '/home/ftp/logs.csv' io.writefile(dst, table.concat(buffer, '\r\n'))

BR,

Erwin

[bmodeco]

Hi Bart,

This script creates a csv of the object logs on your ftp folder:

Code:

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657

-- get past year data logtime = os.time() - 60 * 60 * 24 * 365 -- list of objects by id objects = {} -- objects with logging enabled query = 'SELECT address, datatype, name FROM objects WHERE disablelog=0' for _, object in ipairs(db:getall(query)) do  objects[ tonumber(object.address) ] = {    datatype = tonumber(object.datatype),    name = tostring(object.name or ''),  } end -- csv buffer buffer = { '"date","address","name","value"' } -- get object logs query = 'SELECT src, address, datahex, logtime, eventtype FROM objectlog WHERE logtime >= ? ORDER BY id DESC' for _, row in ipairs(db:getall(query, logtime)) do  object = objects[ tonumber(row.address) ]  -- found matching object and event type is group write  if object and row.eventtype == 'write' then    datatype = object.datatype    -- check that object datatype is set    if datatype then      -- decode data      data = knxdatatype.decode(row.datahex, datatype)      -- remove null chars from char/string datatype      if datatype == dt.char or datatype == dt.string then        data = data:gsub('%z+', '')      -- date to DD.MM.YYYY      elseif datatype == dt.date then        data = string.format('%.2d.%.2d.%.2d', data.day, data.month, data.year)      -- time to HH:MM:SS      elseif datatype == dt.time then        data = string.format('%.2d:%.2d:%.2d', data.hour, data.minute, data.second)      end    else      data = ''    end    -- format csv row    logdate = os.date('%Y.%m.%d %H:%M:%S', row.logtime)    csv = string.format('%q,%q,%q,%q', logdate, knxlib.decodega(row.address), object.name, tostring(data))    -- add to buffer    table.insert(buffer, csv)  end end dst = '/home/ftp/logs.csv' io.writefile(dst, table.concat(buffer, '\r\n'))

BR,

Erwin

Thanks Erwin!

[batistacaceres]

This only work if I have objects, and if objects are mark as log.

If I want to get all traffic in knx without mark the log field, because there is some knx device that i have not control over it, and it send me each time some unknown data, that I have not in the object table.

I just want to send all trafic on KNX bus each 1 hour, even if I have not objects created in my LM5. Could it be posible???

[Erwin van der Zwart]

Hi,

You could try to enable auto object creation in the general settings and set the log policy to all objects.

This way there will be a object created when the telegram holds a valid KNX structure and the logging will be activated already because you log all objects including the new ones..

BR,

Erwin