SpreadsheetDataReader

Short Description

SpreadsheetDataReader reads data from Excel spreadsheets (XLS or XLSX files).

Ports

[1] Records which could not be read correctly are sent to output port 1 if the port has an edge connected. There is a fixed set of fields describing the reason and position of the error which caused the record to fail. Additionally, you can map any field from port 0 as well. Please note: for each error in the input there is one error record generated. That is: for multiple errors in one record you get multiple error records – you can group them, e.g. by the very first integer field.

Processing (parsing) of incorrect or incomplete records depends on error port connection, see Error handling.

This component has metadata templates available.

Metadata

Propagation

SpreadsheetDataReader does not propagate metadata.

Templates

SpreadsheetDataReader has metadata template on error port.

Autofilling Functions

Metadata can use Autofilling Functions.
Note: the source_timestamp and source_size functions work only when reading from a file directly. If the file is an archive or it is stored in a remote location, timestamp will be empty and size will be 0.

SpreadsheetDataReader Attributes

[1] One of these two has to be specified to define the mapping.
[2] Either both or none of these attributes have to be specified.

Details

SpreadsheetDataReader reads data from a specified sheet(s) of XLS or XLSX files. It lets you create complex data mapping: forms, tables, multi-row records, etc.

Supported File Formats

• XLS: only Excel 97/2003 XLS files are supported (BIFF8)
• XLSX: Open Document Format, Microsoft Excel 2007 and newer
  In XLSX, even files with more than 1,048,576 rows can be read although the XLSX format does not officially support it. (Excel will show no more than 2^20 rows.)

Error handling

Processing (parsing) of incorrect or incomplete records depends on error port connection.

• If the error port is not connected, the data parsing stops and next processing is aborted.
• If the error port is connected, record field with an incorrect value or format is sent to error port and data parsing continues.

Introduction to Spreadsheet Mapping

Mapping is a universal pattern guiding the component how to read an Excel spreadsheet. The mapping editor always previews spreadsheets of one file, but mapping can be applied to a whole group of similar files.
Each cell can be mapped to a Clover field in one of the following modes:

• Map by order
  Spreadsheet cells are mapped one by one to output fields in the same order as on the input. If you select another metadata, the cells will be remapped automatically to the new fields.

• Map by name
  Content is mapped to the record field with the same name or label. For each mapped leading cell, the component reads its contents (string) and tries to find a matching field with the same name or label (see Field Name vs. Label vs. Description).
  Fields that could not be mapped to the current file are marked as unresolved. You can either map these explicitly, unmap them or modify output metadata.
  Note that unresolved cells are not a bad thing – for example, you might read a group of similar input files, each containing just a subset of possible columns. Mappings with unresolved cells do not result in your graph failing on execution.

Both Map by order and Map by name modes try to automatically map the contents of the input file to the output metadata. Thus these modes are useful in cases when you read multiple source files and you want to design a single "one-fits-all" generic mapping.

• Explicit
  In Explicit mapping, you explicitly decide which cells are mapped to which record fields.
  This way, you can have, for example, a whole sheet mapped by order with only one cell, which does not fit the mapping, mapped explicitly to a correct field.
  To use explicit mapping, go to Selected cells and fill in Cell value with the target field.
  If a cell is not mapped yet, you might need to switch Mapping mode to Explicit, first. You can also explicitly map a cell to a field by dragging the field from metadata viewer onto the cell. Opposite direction also works (dragging a cell to a field), but you have to first click the cell to select it, because only selected cell can be dragged. Note that you can drag-and-drop more fields/cells at once.

• Auto
  In auto mapping, the first spreadsheet row is whole mapped by name with data offset equal to 1.
  The auto mapping is used, if you leave all the Mapping component property completely blank.
  Another type of auto mapping is created when you map no cell in the mapping editor, but confirm the mapping by clicking the OK button. Then only basic mapping properties will be stored in the Mapping attribute. This way, you can change default Rows per record or Data offset used by the basic implicit mapping mentioned above (if default offset is set to 0, mapping by name is used instead of mapping by name).
  Alternatively, by switching the reading Orientation property, the first column gets implicitly mapped instead of the first row.

Colors in Spreadsheet Mapping Editor

• Orange cells are called leading cells and they form the header. They are a place where a number of mapping settings can be made, see Advanced mapping options below.
• Yellow cells indicate the beginning of the first record.
• Cells in dashed border, which appear after a leading cell is selected, indicate the area data is taken from.

Mapping Editor

The Mapping Editor lets you map spreadsheet rows or columns to metadata fields.
Fill in the File URL and Sheet attributes before opening the Mapping editor. After that, edit Mapping to open a visual mapping editor. It will preview the sheet you have selected:

The Mapping editor consists of these elements:

• Toolbar – buttons controlling how you Map your Excel data (either by order, or by name) and global data offset control (For an explanation of data offsets, see Advanced mapping options below).
• Sheet preview area – this is where you will do and see all the mapping of the source file.
• Output metadata – Clover fields you will map Excel cells to.
• Properties – either for the whole source file (Global) or just the ones concerning Selected cells.
• Summary tab – a place where you can neatly review the whole spreadsheet-to-Data Shaper mapping you have made.

Metadata

Before you start reading a spreadsheet, you might need to extract its metadata as Clover fields (see Extracting Metadata from an XLS(X) File). Note that the extracting wizard resembles the spreadsheet Mapping editor introduced here and it uses the same principle.

You can use the mapping editor to extract metadata right in place without needing to jump to the metadata extract wizard (which is suitable if you need to get just the spreadsheet metadata).

Metadata assigned to the outgoing edge can be edited in the Output metadata area. You can create and manipulate metadata right from the mapping editor, even if you have not connected an output edge (it is created automatically once you create some fields). Available operations include:

• Select existing metadata in the graph using the Output metadata combo.
• Create new metadata using the option in the Output metadata combo.
• Double click a Field to rename it.
• Change data Type via combo-boxes.
• For more operations on the output metadata use the Edit button.
• To create metadata, drag cells from the spreadsheet preview area and drop them between output metadata fields.

Basic Mapping Example

Typically, your Excel data contains headers in the first row and so can be easily mapped. This section describes how to do that.

• First, make sure you have set the Vertical mode in Properties > Global > Orientation. This makes SpreadsheetDataReader process the input by rows (opposite to Horizontal orientation, where reading advances by columns).
• Optional (in case you have not extracted metadata as in Extracting Metadata from an XLS(X) File): select the first row and drag its fields to the Output metadata pane. This will create fields for all cells in the selection. Types will be guessed automatically, but it is worth checking them yourself afterwards.
• Select the whole first row (by clicking the "1" row header) and click either Map by order or Map by name (for explanation, see Introduction to Spreadsheet Mapping above).

• In addition to cell values, it is possible to read other attributes of the input cell: format, formula and hyperlink. See Advanced mapping options below.

Advanced mapping options

This section provides an explanation of some more concepts extending the Basic Mapping Example above.
The first group of options allows reading additional attributes of the selected cell: format, formula and hyperlink.

• Field for format
  Target output field for Excel format pattern (as in Excel’s right-click menu – Format Cells).
  Select a leading cell and set the Field for format property (see figure above) to the target field of type string. You can use this approach even if read data cells have various formats (e.g. different currencies).

If an Excel cell has the General format, the format cannot be transferred to Data Shaper due to internal Excel formatting. Instead, the target field will bear a string "General".

Formats can also be extracted during the one-time metadata extraction process. In metadata, format is taken from a single cell which you supply as a sample value to the metadata extraction wizard. See Extracting Metadata from an XLS(X) File.
If a cell has its format specified by the Excel format string (excel:), SpreadsheetDataReader can read it back. Other readers would ignore it. For further reading on format strings, see Formatting cells (Field with format).

• Field for formula
  Target output field for Excel formulas.
  Select a leading cell and set the Field for formula property (see Advanced Mapping Properties figure above) to the target field of type string.

Reading formulas is only supported in the in-memory mode.

• Field for hyperlink
  Target output field for hyperlink address.
  Select a leading cell and set the Field for hyperlink property (see Advanced Mapping Properties figure above) to the target field of type string..
  Hyperlinks may generally span over multiple cells, but hyperlink reading is only supported for single-cell hyperlinks.

The second group are options that allow more fine-grained control over which cells are mapped to the output record. It is possible to combine multiple source rows into a single output record, specify row offsets or limit the number of output records.

• Data offsets (global)
  Data offsets (global) determines where data is taken from. Basically, its value represents 'a number of rows (in vertical mode) or columns (in horizontal mode) to be omitted - relative to the leading cell (orange)'.
  Click the arrow buttons in the top right corner to adjust data offsets for the whole spreadsheet. Additionally, you can click the spinner in the Selected cells area of each leading cell (the orange one) to adjust data offset locally, i.e. for a particular column only.
  Notice how modifying data offset is visualized in the sheet preview – the 'omitted' rows change color. By following dashed cells, which appear when you click a leading cell, you can quickly state where your record will start at.

Hint!

The arrow buttons in Data offsets (global) only shift the data offset property of each cell either up or down. So mixed offsets are retained, just shifted as desired. To set all data offsets to a single value, enter the value into the number field of Data offsets (global). Note that if there are some mixed offsets, the value is displayed in gray.

• Rows per record
  Rows per record is a Global property specifying how many rows form one record. Best imagined if you look at the figure below:

• Combination of Data offsets and Rows per record
  Combination of Data offsets (global and local) and Rows per record – you can put the settings described in preceding bullet points together. See example:

• Max number of records
  Max number of records is a Global property which you can specify via component attributes, too (see SpreadsheetDataReader Attributes above). If you reduce it, you will notice the number of dashed cells in the spreadsheet preview reduces as well (highlighting only the cells which will be mapped to records in fact).

• Multiple leading cells per column
  In some spreadsheets, data in one column gets mixed, but you still need to process it all into one record. For example, imagine a column containing first names in odd rows and surnames in even rows one after another. In that case, you will create two leading cells above each other to be able to read both first names and surnames. Remember to set Rows per record to an appropriate value (2 in this example) not to read same data in all leading cells. Also, mind raising Data offset in the upper leading cell to start reading data where it truly begins. Look at the figure below:

Notes and Limitations

• Invalid mapping
  It is possible to create invalid mapping using the mapping editor. Invalid mapping causes SpreadsheetDataReader to fail. Such a mapping arises when, for example, one metadata field is mapped to more than one cell, or an autofilled field is mapped (see Autofilling Functions. Another invalid mapping would be caused by an attempt to read a cell (at least one) into more than one metadata field.
  When you change mapping in any way, the validation process is automatically run and you see the warning icon with cell(s) and/or metadata field(s) which cause the mapping to be invalid. When you mouse over such a cell or field, a tooltip with information about the validation problem will be displayed. Also, one of the warning validation messages is displayed at the top of the editor (the white header area).
  Note that warnings caused by cells mapped by name/order will not necessarily lead to the component’s failure (as mentioned earlier).

• Reading formulas
  Reading formulas is only supported in the in-memory mode.

• Reading hyperlinks
  Hyperlinks may generally span over multiple cells, but hyperlink reading is only supported for single-cell hyperlinks. Also, there may be multiple hyperlinks on a particular cell. In such case, one of the hyperlinks will be returned, the order is not specified.

• Reading date as string
  SpreadsheetDataReader cannot guarantee that dates read into string fields will be displayed identically to how they appear in MS Excel. The reason is Data Shaper interprets the format string stored in a cell otherwise than Excel - it depends on your locale.

It is recommended you read dates into date fields and convert them to string using a CTL transformation.

Built-in Excel formats are interpreted according to the following table:

Custom format strings are read as they are defined in Excel. Decimal point is modified according to your locale. Special characters such as double quotes are not interpreted at all.
In both cases (built-in and custom formats), the result may vary from how Excel displays it.

• Reading raw values
  To read numbers from .xls(x) files with full precision available, set the format of the metadata string field to excel:raw. See String Format.
  When SpreadsheetDataReader reads a value from workbook to a string field with the excel:raw format, it tries to extract the raw value from workbook. This has, for example, the following consequences:
• Full precision of numbers is accessible workbook for both: XLS and XLSX.
• Raw representation of dates is available; they are extracted as they are in the workbook - i.e. numbers (both XLS and XLSX)
• Boolean cell values can be retrieved as their xlsx representation - i.e. 0 and 1.

The only exception are shared strings. The referenced string is returned even with the excel:raw format instead of the raw value (the index of the string in the shared strings table).
SpreadsheetDataWriter ignores the excel:raw format. When it is set, the component acts as if the format property is empty.

Examples

Mapping Fields by Order

Read tables with numbers of sold tiles in the first quarter. The tables has the same structure: product name, January, February, March. The company is international. Each affiliate may use a different language, therefore you cannot map fields by name.

Product |January|February|March
T1      |    620|     600|  700
T2      |    150|     150|  100

Producto|Enero  |Febrero |Marzo
T1      |    500|     400|  600
T2      |    300|     400|  500

Solution
Specify the attributes: File URL, Sheet and Mapping.
In Spreadsheet Mapping Editor, map the columns to output metadata fields: select leading cells (the first four cells in the first row) and click Map by order.

See also

SpreadsheetDataWriter
Common Properties of Components
Specific Attribute Types
Common Properties of Readers