(dataserver_data)=
# Data handling

This section describes the data import from an **ASCII** flat file in a {{tds}} of {{uranie}} using 
the formalism of **{{salome}}** tables and JSON format. In both cases, the **ASCII** file is composed 
of 2 parts: the **header** and the **experiment matrix**.

- The **header** describes the information related to the database and related to its attributes:

    - the  name of the database *(optional)*;

    - the title of the database *(optional)*;

    - the date of the database's saving *(optional)*;

    - the name of variables *(mandatory if the title/label's not specified)*. This information will 
    enable to have access to theses variables either to transform them or to produce graphs. The chosen 
    name has to be explicit but rather **short** (smaller than 50 characters);

    - the title/label of variables *(mandatory if the name of variables is not specified)*. This 
    information will be the axis tick marks when this variable is visualised on the plot. As it is 
    possible to use LaTeX commands, the *name* (must be explicit and short) is distinguished from the 
    *label* so as to obtain a good graph rendering.

    - the variable units *(optional)*. This information aims at improving the graph rendering (the 
    unit being displayed next to the label).

    - the nature of the variables *(optional unless dealing with new types)*. This information is not 
    mandatory if one wants to handle only double-precision variables. The possible value being "D" for 
    double, "S" for string and "V" for vectors.

- The second part is the *numerical* experiment matrix.

Using this formalism, it is enough to provide the name of attributes of the ASCII file in order to 
load the file in {{uranie}}.

```{toctree}
data/format_input_output
data/import_from_ascii
data/import_from_TNtupleTDSNtupleTTree
data/add_attributes_tds
data/merge_two_ds
data/pattern_selection
data/export_to_ascii
```