# Creating struct from key-value pairs

**URL:** <https://discuss.hail.is/t/creating-struct-from-key-value-pairs/2049>\
**Category:** Hail Query & hailctl\
**Created:** [May 19, 2021, 8:04pm UTC](https://discuss.hail.is/t/creating-struct-from-key-value-pairs/2049 "2021-05-19T20:04:57Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![hannes\_brt](https://avatars.discourse-cdn.com/v4/letter/h/3be4f8/32.png) [@hannes\_brt](https://discuss.hail.is/u/hannes_brt)\
**Post date:** [May 19, 2021, 8:04pm UTC](https://discuss.hail.is/t/creating-struct-from-key-value-pairs/2049/1 "2021-05-19T20:04:57Z")

</div>

I have a table that contains exon inclusion estimates from RNA-Seq with the columns “event” (exon), “sample” (the individual), “tissue”, and “PSI” (exon inclusion ratio).

I would like to create a matrix table in which the rows are the events, the columns are the samples, and each entry contains a struct where the fields are the tissues and the values are the PSI measurements.

The best I was able to do was `mt = ht.to_matrix_table(["EVENT"], ["sample"])`, but that gives me the following schema:

```auto
----------------------------------------
Global fields:
    None
----------------------------------------
Column fields:
    'sample': str
----------------------------------------
Row fields:
    'EVENT': str
----------------------------------------
Entry fields:
    'PSI': float64
    'tissue': str
----------------------------------------
Column key: ['sample']
Row key: ['EVENT']

```

How can I convert the entry fields to a struct?

Many thanks!

---

<div class="post-metadata">

**Author:** ![danking](https://yyz2.discourse-cdn.com/flex036/user_avatar/discuss.hail.is/danking/32/43_2.png) [@danking](https://discuss.hail.is/u/danking)\
**Post date:** [May 19, 2021, 10:07pm UTC](https://discuss.hail.is/t/creating-struct-from-key-value-pairs/2049/2 "2021-05-19T22:07:23Z")

</div>

`to_matrix_table` assumes a unique mapping from row and column fields to entries, which it shouldn’t (sorry!). For now, you can do this instead:

```auto
In [28]: t = hl.utils.range_table(27) 
    ...:  
    ...: t = t.key_by(event='exon_' + hl.str(t.idx // 9), sample='sample_' + hl.str(t.idx // 3 % 3), tissue='tissue_' + hl.str(t.idx % 3), psi=t.idx).drop('idx') 
    ...:  
    ...: t.show() 
    ...:  
    ...: all_tissues = t.aggregate(hl.agg.collect_as_set(t.tissue)) 
    ...:  
    ...: mt = t.to_matrix_table(['event'], ['sample', 'tissue']) 
    ...: mt = mt.group_cols_by(mt.sample).aggregate(tissue_to_psi_dict = hl.dict(hl.agg.collect((mt.tissue, mt.psi)))) 
    ...: mt = mt.select_entries(**{ 
    ...: tissue_name: mt.tissue_to_psi_dict[tissue_name] for tissue_name in all_tissues 
    ...: }) 
    ...: mt.show() 
    ...:                                                                                                                                                                                       
+----------+------------+------------+-------+
| event | sample | tissue | psi |
+----------+------------+------------+-------+
| str | str | str | int32 |
+----------+------------+------------+-------+
| "exon_0" | "sample_0" | "tissue_0" | 0 |
| "exon_0" | "sample_0" | "tissue_1" | 1 |
| "exon_0" | "sample_0" | "tissue_2" | 2 |
| "exon_0" | "sample_1" | "tissue_0" | 3 |
| "exon_0" | "sample_1" | "tissue_1" | 4 |
| "exon_0" | "sample_1" | "tissue_2" | 5 |
| "exon_0" | "sample_2" | "tissue_0" | 6 |
| "exon_0" | "sample_2" | "tissue_1" | 7 |
| "exon_0" | "sample_2" | "tissue_2" | 8 |
| "exon_1" | "sample_0" | "tissue_0" | 9 |
| "exon_1" | "sample_0" | "tissue_1" | 10 |
| "exon_1" | "sample_0" | "tissue_2" | 11 |
| "exon_1" | "sample_1" | "tissue_0" | 12 |
| "exon_1" | "sample_1" | "tissue_1" | 13 |
| "exon_1" | "sample_1" | "tissue_2" | 14 |
| "exon_1" | "sample_2" | "tissue_0" | 15 |
| "exon_1" | "sample_2" | "tissue_1" | 16 |
| "exon_1" | "sample_2" | "tissue_2" | 17 |
| "exon_2" | "sample_0" | "tissue_0" | 18 |
| "exon_2" | "sample_0" | "tissue_1" | 19 |
| "exon_2" | "sample_0" | "tissue_2" | 20 |
| "exon_2" | "sample_1" | "tissue_0" | 21 |
| "exon_2" | "sample_1" | "tissue_1" | 22 |
| "exon_2" | "sample_1" | "tissue_2" | 23 |
| "exon_2" | "sample_2" | "tissue_0" | 24 |
| "exon_2" | "sample_2" | "tissue_1" | 25 |
| "exon_2" | "sample_2" | "tissue_2" | 26 |
+----------+------------+------------+-------+
2021-05-19 18:06:46 Hail: INFO: Coerced sorted dataset
2021-05-19 18:06:46 Hail: INFO: Coerced dataset with out-of-order partitions.
+----------+---------------------+---------------------+---------------------+---------------------+---------------------+---------------------+---------------------+---------------------+
| event | 'sample_0'.tissue_2 | 'sample_0'.tissue_1 | 'sample_0'.tissue_0 | 'sample_1'.tissue_2 | 'sample_1'.tissue_1 | 'sample_1'.tissue_0 | 'sample_2'.tissue_2 | 'sample_2'.tissue_1 |
+----------+---------------------+---------------------+---------------------+---------------------+---------------------+---------------------+---------------------+---------------------+
| str | int32 | int32 | int32 | int32 | int32 | int32 | int32 | int32 |
+----------+---------------------+---------------------+---------------------+---------------------+---------------------+---------------------+---------------------+---------------------+
| "exon_0" | 2 | 1 | 0 | 5 | 4 | 3 | 8 | 7 |
| "exon_1" | 11 | 10 | 9 | 14 | 13 | 12 | 17 | 16 |
| "exon_2" | 20 | 19 | 18 | 23 | 22 | 21 | 26 | 25 |
+----------+---------------------+---------------------+---------------------+---------------------+---------------------+---------------------+---------------------+---------------------+

+---------------------+
| 'sample_2'.tissue_0 |
+---------------------+
| int32 |
+---------------------+
| 6 |
| 15 |
| 24 |
+---------------------+

```