Qlik sense column consumes memory

[deep2021]

Hi,

In a table, I have a columns containing sting value. The one of the columns is consuming more than 8 gb memory. Is there any approach how we can reduce the size of the column.

Thanks

[Or]

* Reduce the number of distinct values in the column, for example by trimming times from timestamps (or splitting a timestamp into a date and a time component in separate fields). In the case of a string, perhaps your string can be split into parts, with each part having fewer distinct options than the complete string

* Reduce the length of the contents in the column and/or format them explicitly (e.g. num(YourField,#,000)) if the string contains values that can be formatted in this way

* If the value itself isn't necessary and the column is a key field, consider using AutoNumber or a similar function to replace longer contents.

* Remove unnecessary values and replace them with nulls (for example, only keep the last year's worth of strings and not older ones)

You could run QSDA Pro, available here https://easyqlik.com/qsda/download/

It might provide specific recommendations for your column / app.

If none of these options help and the app size is too much of a problem, you could consider using ODAG to only pull part of the data each time rather than having users access the full app.

[rwunderlich]

What is the content of the column?

-Rob

[deep2021]

Hi Rob,

It has paths of documents.

Thanks

[deep2021]

Hi Rob,

Below is the snapshot of data how the path are stored in QS the tables. I have highlighted the columns that are consuming more memory.

Could you please suggest how to reduce the memory consumption for these fields. We need to show these paths on the report.

Thanks,

Deep

[marcus_sommer]

The size of a column respectively a field depends on the number of distinct field-values in it and how numbers are formatted respectively how many chars the (average) strings have. Therefore is it a best practice to split the values from such fields into several fields. Very common is it for example by a timestamp which could easily splitted into a date + time + milliseconds, see: (1) The Importance Of Being Distinct - Qlik Community - 1466796.

In your case the field-values could be also divided, for example within a server + drive + main-folder + sub-folder or similar. With string-functions like substringcount(), subfield() or textbetween() and similar you could create an appropriate logic to split the information into n fields.

- Marcus