(core) Update documentation for lookup/find/prevnext for the Help Center

Summary: Documentation in https://github.com/gristlabs/grist-help/ is built using `./build-functions.sh`, and the current version (as updated and landed in https://github.com/gristlabs/grist-help/pull/368) already reflects the changes in this diff.

Test Plan: No changes to functionality, only documentation comments.

Reviewers: jarek

Reviewed By: jarek

Differential Revision: https://phab.getgrist.com/D4306
This commit is contained in:
Dmitry S 2024-07-31 09:52:02 -07:00
parent 69aabd1ae0
commit 95320f3f5b
3 changed files with 47 additions and 32 deletions

View File

@ -5,24 +5,33 @@ def PREVIOUS(rec, *, group_by=(), order_by):
column IDs, and `order_by` allows column IDs to be prefixed with "-" to reverse sort order. column IDs, and `order_by` allows column IDs to be prefixed with "-" to reverse sort order.
For example, For example,
- `PREVIOUS(rec, order_by="Date")` will return the previous record when the list of records is ```python
sorted by the Date column. PREVIOUS(rec, order_by="Date") # The previous record when sorted by increasing Date.
- `PREVIOUS(rec, order_by="-Date")` will return the previous record when the list is sorted by PREVIOUS(rec, order_by="-Date") # The previous record when sorted by decreasing Date.
the Date column in descending order. ```
- `PREVIOUS(rec, group_by="Account", order_by="Date")` will return the previous record with the
same Account as `rec`, when records are filtered by the Account of `rec` and sorted by Date. You may use `group_by` to search for the previous record within a filtered group. For example,
this finds the previous record with the same Account as `rec`, when records are filtered by the
Account of `rec` and sorted by increasing Date:
```python
PREVIOUS(rec, group_by="Account", order_by="Date")
```
When multiple records have the same `order_by` values (e.g. the same Date in the examples above), When multiple records have the same `order_by` values (e.g. the same Date in the examples above),
the order is determined by the relative position of rows in views. This is done internally by the order is determined by the relative position of rows in views. This is done internally by
falling back to the special column `manualSort` and the row ID column `id`. falling back to the special column `manualSort` and the row ID column `id`.
Use `order_by=None` to find the previous record in an unsorted table (when rows may be Use `order_by=None` to find the previous record in an unsorted table (when rows may be
rearranged by dragging them manually). For example, rearranged by dragging them manually). For example:
- `PREVIOUS(rec, order_by=None)` will return the previous record in the unsorted list of records. ```python
PREVIOUS(rec, order_by=None) # The previous record in the unsorted list of records.
```
You may specify multiple column IDs as a tuple, for both `group_by` and `order_by`. This can be You may specify multiple column IDs as a tuple, for both `group_by` and `order_by`. This can be
used to match views sorted by multiple columns. For example: used to match views sorted by multiple columns. For example:
- `PREVIOUS(rec, group_by=("Account", "Year"), order_by=("Date", "-Amount"))` ```python
PREVIOUS(rec, group_by=("Account", "Year"), order_by=("Date", "-Amount"))
```
""" """
return _sorted_lookup(rec, group_by=group_by, order_by=order_by)._find.previous(rec) return _sorted_lookup(rec, group_by=group_by, order_by=order_by)._find.previous(rec)
@ -39,10 +48,10 @@ def RANK(rec, *, group_by=(), order_by, order="asc"):
`order_by`, and grouping specified by `group_by`. See [`PREVIOUS`](#previous) for details of `order_by`, and grouping specified by `group_by`. See [`PREVIOUS`](#previous) for details of
these parameters. these parameters.
The `order` parameter may be "asc" (which is the default) or "desc". The `order` parameter may be `"asc"` (which is the default) or `"desc"`.
When `order` is "asc" or omitted, the first record in the group in the sorted order would have When `order` is `"asc"` or omitted, the first record in the group in the sorted order would have
the rank of 1. When `order` is "desc", the last record in the sorted order would have the rank the rank of 1. When `order` is `"desc"`, the last record in the sorted order would have the rank
of 1. of 1.
If there are multiple groups, there will be multiple records with the same rank. In particular, If there are multiple groups, there will be multiple records with the same rank. In particular,
@ -50,7 +59,7 @@ def RANK(rec, *, group_by=(), order_by, order="asc"):
For example, `RANK(rec, group_by="Year", order_by="Score", order="desc")` will return the rank of For example, `RANK(rec, group_by="Year", order_by="Score", order="desc")` will return the rank of
the current record (`rec`) among all the records in its table for the same year, ordered by the current record (`rec`) among all the records in its table for the same year, ordered by
score. decreasing score.
""" """
return _sorted_lookup(rec, group_by=group_by, order_by=order_by)._find.rank(rec, order=order) return _sorted_lookup(rec, group_by=group_by, order_by=order_by)._find.rank(rec, order=order)

View File

@ -243,26 +243,30 @@ class RecordSet(object):
@property @property
def find(self): def find(self):
""" """
A set of methods for finding values in sorted set of records. For example: Name: find.*
Usage: RecordSet.**find.\\***(value)
A set of methods for finding values in sorted sets of records, as returned by
[`lookupRecords`](#lookuprecords). For example:
``` ```
Transactions.lookupRecords(..., sort_by="Date").find.lt($Date) Transactions.lookupRecords(..., order_by="Date").find.lt($Date)
Table.lookupRecords(..., sort_by=("Foo", "Bar")).find.le(foo, bar) Table.lookupRecords(..., order_by=("Foo", "Bar")).find.le(foo, bar)
``` ```
If the `find` method is shadowed by a same-named user column, you may use `_find` instead. If the `find` attribute is shadowed by a same-named user column, you may use `_find` instead.
The methods available are: The methods available are:
- `lt`: (less than) find nearest record with sort values < the given values - __`lt`__: (less than) find nearest record with sort values < the given values
- `le`: (less than or equal to) find nearest record with sort values <= the given values - __`le`__: (less than or equal to) find nearest record with sort values <= the given values
- `gt`: (greater than) find nearest record with sort values > the given values - __`gt`__: (greater than) find nearest record with sort values > the given values
- `ge`: (greater than or equal to) find nearest record with sort values >= the given values - __`ge`__: (greater than or equal to) find nearest record with sort values >= the given values
- `eq`: (equal to) find nearest record with sort values == the given values - __`eq`__: (equal to) find nearest record with sort values == the given values
Example from https://templates.getgrist.com/5pHLanQNThxk/Payroll. Each person has a history of Example from [our Payroll template](https://templates.getgrist.com/5pHLanQNThxk/Payroll).
pay rates, in the Rates table. To find a rate applicable on a certain date, here is how you Each person has a history of pay rates, in the Rates table. To find a rate applicable on a
can do it old-style: certain date, here is how you can do it old-style:
``` ```python
# Get all the rates for the Person and Role in this row. # Get all the rates for the Person and Role in this row.
rates = Rates.lookupRecords(Person=$Person, Role=$Role) rates = Rates.lookupRecords(Person=$Person, Role=$Role)
@ -277,8 +281,9 @@ class RecordSet(object):
``` ```
With the new methods, it is much simpler: With the new methods, it is much simpler:
``` ```python
rate = Rates.lookupRecords(Person=$Person, Role=$Role, sort_by="Rate_Start").find.le($Date) rates = Rates.lookupRecords(Person=$Person, Role=$Role, order_by="Rate_Start")
rate = rates.find.le($Date)
return rate.Hourly_Rate return rate.Hourly_Rate
``` ```

View File

@ -95,7 +95,8 @@ class UserTable(object):
For backward compatibility, `sort_by` may be used instead of `order_by`, but only allows a For backward compatibility, `sort_by` may be used instead of `order_by`, but only allows a
single field, and falls back to row ID (rather than `manualSort`). single field, and falls back to row ID (rather than `manualSort`).
See [RecordSet](#recordset) for useful properties offered by the returned object. See [RecordSet](#recordset) for useful properties offered by the returned object. In
particular, methods like [`.find.le`](#find_) allow searching for nearest values.
See [CONTAINS](#contains) for an example utilizing `UserTable.lookupRecords` to find records See [CONTAINS](#contains) for an example utilizing `UserTable.lookupRecords` to find records
where a field of a list type (such as `Choice List` or `Reference List`) contains the given where a field of a list type (such as `Choice List` or `Reference List`) contains the given
@ -127,13 +128,13 @@ class UserTable(object):
parameter to the column ID by which to sort the matches, to determine which of them is parameter to the column ID by which to sort the matches, to determine which of them is
returned as the first one. By default, the record with the lowest row ID is returned. returned as the first one. By default, the record with the lowest row ID is returned.
See [`lookupRecords`](#lookupRecords) for details of all available options and behavior of See [`lookupRecords`](#lookuprecords) for details of all available options and behavior of
`order_by` (and of its legacy alternative, `sort_by`). `order_by` (and of its legacy alternative, `sort_by`).
For example: For example:
``` ```
Tasks.lookupOne(Project=$id, order_by="Priority") # Returns the Task with the smallest Priority. Tasks.lookupOne(Project=$id, order_by="Priority") # Task with the smallest Priority.
Rates.lookupOne(Person=$id, order_by="-Date") # Returns the Rate with the latest Date. Rates.lookupOne(Person=$id, order_by="-Date") # Rate with the latest Date.
``` ```
""" """
return self.table.lookup_one_record(**field_value_pairs) return self.table.lookup_one_record(**field_value_pairs)